{"aliases": [], "authors": ["Panupong Pasupat", "Tian-Shun Jiang", "Evan Zheran Liu", "Kelvin Guu", "Percy Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f368116f70e987cde111835a05b5b36e0d345dabb4e8351610e2bbfb691b3440", "submitted_at": "2018-08-28T06:09:39Z", "updated_at": "2018-10-01T03:22:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1808.09132v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:1808.09132", "identifiers": {"arxiv": "1808.09132"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1808.09132"}], "name": "Mapping Natural Language Commands to Web Elements", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2018-08-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Defines language grounding on webpages as choosing the UI element targeted by a natural-language command. A collection of more than 50,000 commands and three baseline models study functional references, relationships between elements, and visual position cues.", "topics": ["grounding", "data-generation", "evaluation"], "updated_at": "2018-10-01", "url": "https://arxiv.org/abs/1808.09132"}
{"aliases": [], "authors": ["Yang Li", "Jiacong He", "Xin Zhou", "Yuan Zhang", "Jason Baldridge"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3af7b45455ffbd616e8cdf5612b19275ecb7bf07b1065af9f5b724ae19e96fb4", "submitted_at": "2020-05-07T21:41:40Z", "updated_at": "2020-06-05T02:11:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2005.03776v2"}, "publication": {"authors": ["Yang Li", "Jiacong He", "Xin Zhou", "Yuan Zhang", "Jason Baldridge"], "date_basis": "Explicit publisher metadata: 2020/7; precision=month.", "date_parts": [2020, 7], "doi": "10.18653/v1/2020.acl-main.729", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2020.acl-main.729/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "8198-8210", "pdf_url": "https://aclanthology.org/2020.acl-main.729.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Mapping Natural Language Instructions to Mobile UI Action Sequences", "type": "conference-paper", "url": "https://aclanthology.org/2020.acl-main.729/", "venue": "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "c196294e19378d08d71a7cfd41c2e368f5a6b49ee7c3601257537b9ae5c42add", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2020.acl-main.729/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2005.03776", "identifiers": {"arxiv": "2005.03776"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2005.03776"}, {"rel": "publication", "url": "https://aclanthology.org/2020.acl-main.729/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2020.acl-main.729.pdf"}], "name": "Mapping Natural Language Instructions to Mobile UI Action Sequences", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2020-05-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PIXELHELP for predicting mobile UI action sequences from instructions, together with action-phrase annotations and synthetic UI action descriptions for training. Separate Transformers extract instruction steps and ground object descriptions using interface content and screen position.", "topics": ["mobile-navigation", "grounding", "data-generation", "action-prediction"], "updated_at": "2020-06-05", "url": "https://arxiv.org/abs/2005.03776"}
{"aliases": [], "authors": ["Jieshan Chen", "Mulong Xie", "Zhenchang Xing", "Chunyang Chen", "Xiwei Xu", "Liming Zhu", "Guoqiang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6352985532945fd562ce0382aef0e977ce1ec5909b6872e5fe048167e9732a0", "submitted_at": "2020-08-12T06:36:33Z", "updated_at": "2020-09-07T12:57:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2008.05132v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2008.05132", "identifiers": {"arxiv": "2008.05132"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2008.05132"}], "name": "Object Detection for Graphical User Interface: Old Fashioned or Deep Learning or a Combination?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2020-08-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares seven traditional and learned GUI element detectors on more than 50,000 interface images. The analysis motivates a coarse-to-fine non-text detector combined with a learned text detector to address GUI-specific localization requirements.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2020-09-07", "url": "https://arxiv.org/abs/2008.05132"}
{"aliases": [], "authors": ["Yang Li", "Gang Li", "Luheng He", "Jingjie Zheng", "Hong Li", "Zhiwei Guan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68af0a0a4b605d0c968e8913232fde16e3eadf6361099e8fc3d015dcdb26302b", "submitted_at": "2020-10-08T22:56:03Z", "updated_at": "2020-10-08T22:56:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2010.04295v1"}, "publication": {"authors": ["Yang Li", "Gang Li", "Luheng He", "Jingjie Zheng", "Hong Li", "Zhiwei Guan"], "date_basis": "Explicit publisher metadata: 2020/11; precision=month.", "date_parts": [2020, 11], "doi": "10.18653/v1/2020.emnlp-main.443", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2020.emnlp-main.443/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The retained arXiv abstract reports 162,859 phrases; the ACL abstract reports 162,860. This publication proposal does not overwrite the arXiv annotation or assert a unified dataset count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "5495-5510", "pdf_url": "https://aclanthology.org/2020.emnlp-main.443.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements", "type": "conference-paper", "url": "https://aclanthology.org/2020.emnlp-main.443/", "venue": "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "750e60b4d28e5b97a46595ea2280b520bf9edcc445b9931bad225e0028935c5e", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2020.emnlp-main.443/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2010.04295", "identifiers": {"arxiv": "2010.04295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2010.04295"}, {"rel": "publication", "url": "https://aclanthology.org/2020.emnlp-main.443/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2020.emnlp-main.443.pdf"}], "name": "Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2020-10-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces crowdsourced descriptions of mobile UI elements and studies multimodal models that generate useful widget captions.", "topics": ["screen-understanding", "grounding"], "updated_at": "2020-10-08", "url": "https://arxiv.org/abs/2010.04295"}
{"aliases": [], "authors": ["Zecheng He", "Srinivas Sunkara", "Xiaoxue Zang", "Ying Xu", "Lijuan Liu", "Nevan Wichers", "Gabriel Schubiner", "Ruby Lee", "Jindong Chen", "Blaise Agüera y Arcas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51fdddb9a3d8a6090875719a51b22008b1f43bc38c6cfe1713a49a52f7e8434c", "submitted_at": "2020-12-22T20:49:52Z", "updated_at": "2021-01-25T20:37:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2012.12350v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "139375513dafe5d7cfd0f5843b2a3465294e2b54051cf1c964a0d0751166a469", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2012.12350v2"}], "id": "arxiv:2012.12350", "identifiers": {"arxiv": "2012.12350"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2012.12350"}], "name": "ActionBert: Leveraging User Actions for Semantic Understanding of User Interfaces", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2020-12-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Pretrains ActionBert on user interaction traces to combine visual, linguistic, and UI-structure features. It evaluates the resulting UI representations on tasks including icon classification and description-based component retrieval.", "topics": ["screen-understanding", "grounding", "action-prediction"], "updated_at": "2021-01-25", "url": "https://arxiv.org/abs/2012.12350"}
{"aliases": [], "authors": ["Nancy Xu", "Sam Masling", "Michael Du", "Giovanni Campagna", "Larry Heck", "James Landay", "Monica S Lam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e00b8f1b34097455ece2cf97efabbb64f398e3fffdec13394f731096672dc654", "submitted_at": "2021-03-30T04:02:34Z", "updated_at": "2021-04-04T23:51:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2103.16057v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2103.16057", "identifiers": {"arxiv": "2103.16057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2103.16057"}], "name": "Grounding Open-Domain Instructions to Automate Web Support Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-03-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces RUSS, which translates support instructions into a domain-specific language and grounds them in actionable web elements.", "topics": ["grounding", "web-navigation"], "updated_at": "2021-04-04", "url": "https://arxiv.org/abs/2103.16057"}
{"aliases": [], "authors": ["Jingwen Fu", "Xiaoyi Zhang", "Yuwang Wang", "Wenjun Zeng", "Sam Yang", "Grayson Hilliard"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8788e419d1a901f8ad9dd6089a12fe129baee764b1c285bbeb4ea60d7046afa2", "submitted_at": "2021-05-25T13:45:54Z", "updated_at": "2022-04-14T11:00:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2105.11941v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "41c4da71e457ea23725a8070b0da64c234f03613841b95edea3a5f3c95dd8c68", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2105.11941v2"}], "id": "arxiv:2105.11941", "identifiers": {"arxiv": "2105.11941"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2105.11941"}], "name": "Understanding Mobile GUI: from Pixel-Words to Screen-Sentences", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2021-05-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Models mobile screens as atomic text and graphic components whose relationships are learned by a Screen Transformer. The work introduces RICO-PW annotations and evaluates metadata-free screen understanding on relation, clickability, retrieval, and app-classification tasks.", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2022-04-14", "url": "https://arxiv.org/abs/2105.11941"}
{"aliases": [], "authors": ["Yang Li", "Gang Li", "Xin Zhou", "Mostafa Dehghani", "Alexey Gritsenko"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e88e6e2659df232174d9fc71a2fac61a021941f096355f8ea3cd4c618c223d3", "submitted_at": "2021-12-10T17:37:26Z", "updated_at": "2021-12-10T17:37:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.05692v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2112.05692", "identifiers": {"arxiv": "2112.05692"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2112.05692"}], "name": "VUT: Versatile UI Transformer for Multi-Modal Multi-Task User Interface Modeling", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-12-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "The Versatile UI Transformer jointly models interface screenshots, structure, and language for five UI tasks. Its encoder supports element detection when structural metadata is absent, while an autoregressive component handles UI question answering and command grounding.", "topics": ["screen-understanding", "grounding", "efficiency"], "updated_at": "2021-12-10", "url": "https://arxiv.org/abs/2112.05692"}
{"aliases": [], "authors": ["Gang Li", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eaa351e34fa6bdcbe26aa6714eacda1f667411090abfdd5575735605dc89e864", "submitted_at": "2022-09-29T16:45:43Z", "updated_at": "2023-02-24T01:41:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2209.14927v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2209.14927", "identifiers": {"arxiv": "2209.14927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.14927"}], "name": "Spotlight: Mobile UI Understanding using Vision-Language Models with a Focus", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2022-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Spotlight, a vision-language model conditioned on a UI screenshot and a selected region of interest. It evaluates screenshot-based mobile UI understanding and explores multitask learning and few-shot prompting.", "topics": ["screen-understanding", "grounding"], "updated_at": "2023-02-24", "url": "https://arxiv.org/abs/2209.14927"}
{"aliases": [], "authors": ["Tao Li", "Gang Li", "Jingjie Zheng", "Purple Wang", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15e57e7dd30e6f3ae8b13cbd4a190b39339f2bbf7a551a6a5df5190dc62b5434", "submitted_at": "2022-09-29T21:08:18Z", "updated_at": "2022-09-29T21:08:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2209.15099v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2209.15099", "identifiers": {"arxiv": "2209.15099"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.15099"}], "name": "MUG: Interactive Multimodal Grounding on User Interfaces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies iterative UI grounding in which users refine or correct an agent's interpretation over multiple conversational turns.", "topics": ["grounding", "human-in-the-loop"], "updated_at": "2022-09-29", "url": "https://arxiv.org/abs/2209.15099"}
{"aliases": [], "authors": ["Srinivas Sunkara", "Maria Wang", "Lijuan Liu", "Gilles Baechler", "Yu-Chung Hsiao", "Jindong", "Chen", "Abhanshu Sharma", "James Stout"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "24d037f1abeb7708852ac5517ce02cc46622e359d9feab407f188c0657cc72f4", "submitted_at": "2022-10-06T03:48:54Z", "updated_at": "2022-10-06T03:48:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2210.02663v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2210.02663", "identifiers": {"arxiv": "2210.02663"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2210.02663"}], "name": "Towards Better Semantic Understanding of Mobile Interfaces", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2022-10-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Augments RICO screens and view hierarchies with human labels for icon shape and meaning and for links between UI elements and text labels. It compares image-only and multimodal models on semantic understanding of previously unseen apps.", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2022-10-06", "url": "https://arxiv.org/abs/2210.02663"}
{"aliases": [], "authors": ["Sagar Gubbi Venkatesh", "Partha Talukdar", "Srini Narayanan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52319819402cec76a69c0bb513f3a7000053208ddd9d2a8c675e055422402fbe", "submitted_at": "2022-11-14T18:36:19Z", "updated_at": "2023-05-23T16:08:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2211.07615v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2211.07615", "identifiers": {"arxiv": "2211.07615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2211.07615"}], "name": "UGIF: UI Grounded Instruction Following", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-11-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces multilingual smartphone instruction-following data and a pipeline that retrieves help instructions and maps them to UI actions.", "topics": ["mobile-navigation", "grounding"], "updated_at": "2023-05-23", "url": "https://arxiv.org/abs/2211.07615"}
{"aliases": [], "authors": ["Pratyay Banerjee", "Shweti Mahajan", "Kushal Arora", "Chitta Baral", "Oriana Riva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb647a96c7f8263786e85a5061c8c5284cb92ed0b2ebc4c405e1243b2ca3fa2b", "submitted_at": "2023-01-23T09:05:49Z", "updated_at": "2023-01-23T09:05:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2301.10165v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2301.10165", "identifiers": {"arxiv": "2301.10165"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2301.10165"}], "name": "Lexi: Self-Supervised Learning of the UI Language", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2023-01-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Lexi and UICaption to learn UI image-text representations from descriptions of interface functionality. Evaluation covers UI action entailment, instruction-based retrieval, referring-expression grounding, and UI entity recognition.", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2023-01-23", "url": "https://arxiv.org/abs/2301.10165"}
{"aliases": [], "authors": ["Peter Shaw", "Mandar Joshi", "James Cohan", "Jonathan Berant", "Panupong Pasupat", "Hexiang Hu", "Urvashi Khandelwal", "Kenton Lee", "Kristina Toutanova"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d06f7480695f45546b62b5b8c3d8c174ca52936e158d46e626689d0918711417", "submitted_at": "2023-05-31T23:39:18Z", "updated_at": "2023-12-06T23:46:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.00245v2"}, "publication": {"authors": ["Peter Shaw", "Mandar Joshi", "James Cohan", "Jonathan Berant", "Panupong Pasupat", "Hexiang Hu", "Urvashi Khandelwal", "Kenton Lee", "Kristina Toutanova"], "date_basis": "Explicit publisher metadata: 2023-12-15; precision=day.", "date_parts": [2023, 12, 15], "doi": "10.52202/075280-1490", "event_dates": null, "evidence_urls": ["https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html", "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The NeurIPS landing metadata adds the initial N to Kristina Toutanova; the formal paper title page does not. Publication authors use the PDF. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "34354-34370", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf", "publisher": null, "series": null, "title": "From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces", "type": "conference-paper", "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "36"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "938e0a31512bb9a7793b568858f81eb3ceb650a71aae2d91b5292990e890201b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:47+00:00", "role": "primary", "sha256": "3827298f44d2f24f1333b4d46ecb5eb88c8b028d192b11424580655b38e024f9", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2306.00245", "identifiers": {"arxiv": "2306.00245"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.00245"}, {"rel": "publication", "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"}], "name": "From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2023-05-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies instruction-following agents that observe screenshots and act through a generic mouse-and-keyboard action space, evaluating pixel-based pretraining for GUI tasks on MiniWoB++.", "topics": ["grounding", "action-prediction", "web-navigation"], "updated_at": "2023-12-06", "url": "https://arxiv.org/abs/2306.00245"}
{"aliases": [], "authors": ["Zhizheng Zhang", "Wenxuan Xie", "Xiaoyi Zhang", "Yan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f389de80bb86a759df2cfc7000188286efcd73c99a43ede376c46f190832ef80", "submitted_at": "2023-10-07T07:22:41Z", "updated_at": "2023-10-07T07:22:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.04716v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.04716", "identifiers": {"arxiv": "2310.04716"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.04716"}], "name": "Reinforced UI Instruction Grounding: Towards a Generic UI Task Automation API", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains a screenshot-based instruction-grounding model to predict target coordinates, using reinforcement learning to improve spatial decoding.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2023-10-07", "url": "https://arxiv.org/abs/2310.04716"}
{"aliases": [], "authors": ["Yue Jiang", "Eldon Schoop", "Amanda Swearngin", "Jeffrey Nichols"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a49e879891017071fce4df0ff4b07bd9e8b5a6e14385af565549aadbde1bc1cf", "submitted_at": "2023-10-07T16:32:34Z", "updated_at": "2023-10-07T16:32:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.04869v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.04869", "identifiers": {"arxiv": "2310.04869"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.04869"}], "name": "ILuvUI: Instruction-tuned LangUage-Vision modeling of UIs from Machine Conversations", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "ILuvUI generates 335,000 conversational training examples paired with UI screenshots, covering questions, descriptions, and plans. The resulting vision-language model is assessed on element detection, response quality, and examples of multi-step UI navigation.", "topics": ["data-generation", "screen-understanding", "grounding", "planning"], "updated_at": "2023-10-07", "url": "https://arxiv.org/abs/2310.04869"}
{"aliases": [], "authors": ["Jianwei Yang", "Hao Zhang", "Feng Li", "Xueyan Zou", "Chunyuan Li", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "abafe2b974ddd80416b215cd4c48d801e18fe9162b9c097d819d42eaff68a069", "submitted_at": "2023-10-17T17:51:31Z", "updated_at": "2023-11-06T07:39:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.11441v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.11441", "identifiers": {"arxiv": "2310.11441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.11441"}], "name": "Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4V", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-17", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Overlays labeled image regions to help multimodal models refer to and reason about visual targets; a general visual-prompting method relevant to GUI grounding.", "topics": ["grounding"], "updated_at": "2023-11-06", "url": "https://arxiv.org/abs/2310.11441"}
{"aliases": [], "authors": ["An Yan", "Zhengyuan Yang", "Wanrong Zhu", "Kevin Lin", "Linjie Li", "Jianfeng Wang", "Jianwei Yang", "Yiwu Zhong", "Julian McAuley", "Jianfeng Gao", "Zicheng Liu", "Lijuan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebe6ff09f876b81d033a064301a1c23e3c141d2a1ad0d7a5df91e729377fcd3b", "submitted_at": "2023-11-13T18:53:37Z", "updated_at": "2023-11-13T18:53:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.07562v1"}}, "environments": ["ios", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2311.07562", "identifiers": {"arxiv": "2311.07562"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.07562"}], "name": "GPT-4V in Wonderland: Large Multimodal Models for Zero-Shot Smartphone GUI Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-11-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies MM-Navigator, a multimodal agent that interprets smartphone screenshots and predicts localized actions without task-specific training.", "topics": ["mobile-navigation", "grounding"], "updated_at": "2023-11-13", "url": "https://arxiv.org/abs/2311.07562"}
{"aliases": [], "authors": ["Wenyi Hong", "Weihan Wang", "Qingsong Lv", "Jiazheng Xu", "Wenmeng Yu", "Junhui Ji", "Yan Wang", "Zihan Wang", "Yuxuan Zhang", "Juanzi Li", "Bin Xu", "Yuxiao Dong", "Ming Ding", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755198d83a3af9fe6d032f1446ec9bf57bf7e6f314d37787ce8a6aaa16bddc47", "submitted_at": "2023-12-14T13:20:57Z", "updated_at": "2024-12-27T06:56:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.08914v3"}, "publication": {"authors": ["Wenyi Hong", "Weihan Wang", "Qingsong Lv", "Jiazheng Xu", "Wenmeng Yu", "Junhui Ji", "Yan Wang", "Zihan Wang", "Yuxiao Dong", "Ming Ding", "Jie Tang"], "date_basis": "Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.", "date_parts": [2024, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html", "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher page directly links this arXiv identifier. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, and landing page list eleven authors; the retained arXiv snapshot lists fourteen. Each edition retains its own authors.", "pages": "14281-14290", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf", "publisher": null, "series": null, "title": "CogAgent: A Visual Language Model for GUI Agents", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition", "volume": null}}, "environments": ["desktop", "android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:00Z", "role": "primary", "sha256": "43009fbc322dbe4f6cd1be01b7f414b4747ddcadaf4b63ae16f8eecc7ca6837b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:11.606232+00:00", "role": "primary", "sha256": "b9bfea178ee72b694eb575c73e3d7d8209a72706c9dda04296a70e17b0c98cf4", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2312.08914", "identifiers": {"arxiv": "2312.08914"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.08914"}, {"rel": "publication", "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html"}, {"rel": "publication_pdf", "url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"}], "name": "CogAgent: A Visual Language Model for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-12-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a visual-language model with low- and high-resolution image encoders for screenshot understanding and GUI navigation.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-12-27", "url": "https://arxiv.org/abs/2312.08914"}
{"aliases": [], "authors": ["Boyuan Zheng", "Boyu Gou", "Jihyung Kil", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a005fb14b21cf17521f4ea966e555d9963092b709bd9280c6e72a9bf0f5fa93a", "submitted_at": "2024-01-03T08:33:09Z", "updated_at": "2024-03-12T23:14:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.01614v2"}, "publication": {"authors": ["Boyuan Zheng", "Boyu Gou", "Jihyung Kil", "Huan Sun", "Yu Su"], "date_basis": "Explicit publisher metadata: 2024/07/08; precision=day.", "date_parts": [2024, 7, 8], "doi": null, "event_dates": "21--27 Jul 2024", "evidence_urls": ["https://proceedings.mlr.press/v235/zheng24e.html"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "61349-61385", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/zheng24e/zheng24e.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "GPT-4V(ision) is a Generalist Web Agent, if Grounded", "type": "conference-paper", "url": "https://proceedings.mlr.press/v235/zheng24e.html", "venue": "Proceedings of the 41st International Conference on Machine Learning", "volume": "235"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "eb144b5dfe51e5a9092d273fcffb85343808d011b03d22e059c9cc65e642ef33", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v235/zheng24e.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2401.01614", "identifiers": {"arxiv": "2401.01614"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.01614"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v235/zheng24e.html"}, {"rel": "publication_pdf", "url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/zheng24e/zheng24e.pdf"}], "name": "GPT-4V(ision) is a Generalist Web Agent, if Grounded", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-01-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SeeAct and live-website evaluation, examining how visual planning and element grounding affect web-agent task completion.", "topics": ["web-navigation", "grounding"], "updated_at": "2024-03-12", "url": "https://arxiv.org/abs/2401.01614"}
{"aliases": [], "authors": ["Kanzhi Cheng", "Qiushi Sun", "Yougang Chu", "Fangzhi Xu", "Yantao Li", "Jianbing Zhang", "Zhiyong Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c92ec9323eec9362aa07984f9fb1cbfbe98ecae426a9f8adc9d436328a87aed3", "submitted_at": "2024-01-17T08:10:35Z", "updated_at": "2024-02-23T04:36:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.10935v2"}, "publication": {"authors": ["Kanzhi Cheng", "Qiushi Sun", "Yougang Chu", "Fangzhi Xu", "Yantao Li", "Jianbing Zhang", "Zhiyong Wu"], "date_basis": "Explicit publisher metadata: 2024/8; precision=month.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.505", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.505/", "https://aclanthology.org/2024.acl-long.505.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The ACL landing metadata spells the fifth author Li YanTao. The formal PDF title page spells Yantao Li; the publication author field uses the PDF. The retained arXiv author list is untouched. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "9313-9332", "pdf_url": "https://aclanthology.org/2024.acl-long.505.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.505/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "62d56e417487461c8b34da15167eafd5f14f58092e4269222dc5c1c9048731a9", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.505/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "8e1f6809e35751d762714286c55e79e81bae071608f90ef561d804aa59d7e406", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2024.acl-long.505.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2401.10935", "identifiers": {"arxiv": "2401.10935"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.10935"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.505/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.505.pdf"}], "name": "SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-01-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains a screenshot-driven GUI agent with grounding pretraining and introduces ScreenSpot for evaluating interface target localization.", "topics": ["grounding", "data-generation"], "updated_at": "2024-02-23", "url": "https://arxiv.org/abs/2401.10935"}
{"aliases": [], "authors": ["Jing Yu Koh", "Robert Lo", "Lawrence Jang", "Vikram Duvvur", "Ming Chong Lim", "Po-Yu Huang", "Graham Neubig", "Shuyan Zhou", "Ruslan Salakhutdinov", "Daniel Fried"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a705b87b43464120ebb01b28b0f6a9e31fad68f3d463c88354f2005dab940c3", "submitted_at": "2024-01-24T18:35:21Z", "updated_at": "2024-06-06T02:01:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.13649v2"}, "publication": {"authors": ["Jing Yu Koh", "Robert Lo", "Lawrence Jang", "Vikram Duvvur", "Ming Chong Lim", "Po-Yu Huang", "Graham Neubig", "Shuyan Zhou", "Ruslan Salakhutdinov", "Daniel Fried"], "date_basis": "Explicit publisher metadata: 2024/8; precision=month.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.50", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.50/", "https://aclanthology.org/2024.acl-long.50.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The formal PDF title adds Visually Grounded; the landing metadata uses Visual Web Tasks. PDF authors include Ming Chong Lim and Ruslan Salakhutdinov; the landing abbreviates them to Ming Lim and Russ Salakhutdinov. Publication title/authors use the PDF; the retained arXiv title remains unchanged. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "881-905", "pdf_url": "https://aclanthology.org/2024.acl-long.50.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "VisualWebArena: Evaluating Multimodal Agents on Realistic Visually Grounded Web Tasks", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.50/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "fc908d474b12a6c0f3ff3e48129d29c809e9f3a0365a3cf647b70273f266849a", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.50/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "6fc85c38c975c2af170c1d1a28f72ba557906d9019d1d9106a15215868bb491e", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://aclanthology.org/2024.acl-long.50.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2401.13649", "identifiers": {"arxiv": "2401.13649"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.13649"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.50/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.50.pdf"}], "name": "VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web Tasks", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-01-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces visually grounded website tasks for evaluating multimodal agents that must interpret both images and text.", "topics": ["evaluation", "grounding", "web-navigation"], "updated_at": "2024-06-06", "url": "https://arxiv.org/abs/2401.13649"}
{"aliases": [], "authors": ["Jihyung Kil", "Chan Hee Song", "Boyuan Zheng", "Xiang Deng", "Yu Su", "Wei-Lun Chao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3da421fc440f7dcfc53d8ef218f55389e4da4dbac196f54b40402bc877ff0365", "submitted_at": "2024-02-06T23:52:10Z", "updated_at": "2024-03-30T05:18:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.04476v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.04476", "identifiers": {"arxiv": "2402.04476"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.04476"}], "name": "Dual-View Visual Contextualization for Web Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses screenshot geometry and visual content to enrich HTML element representations for web navigation. Nearby elements contribute textual and visual context, with evaluation on cross-task, cross-website, and cross-domain Mind2Web splits.", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2024-03-30", "url": "https://arxiv.org/abs/2402.04476"}
{"aliases": [], "authors": ["Chaoyun Zhang", "Liqun Li", "Shilin He", "Xu Zhang", "Bo Qiao", "Si Qin", "Minghua Ma", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "533acb1c9aa3f1c77adb2c67779da86e3d5de95264ba3ec1f81d803f67da7753", "submitted_at": "2024-02-08T15:40:35Z", "updated_at": "2024-05-23T05:18:58Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2402.07939v5"}, "publication": {"authors": ["Chaoyun Zhang", "Liqun Li", "Shilin He", "Xu Zhang", "Bo Qiao", "Si Qin", "Minghua Ma", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.26", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.26/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "597-622", "pdf_url": "https://aclanthology.org/2025.naacl-long.26.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "UFO: A UI-Focused Agent for Windows OS Interaction", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.26/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}}, "environments": ["windows", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:48Z", "role": "primary", "sha256": "bbe15e312bcbc923811ef73e561fa7db6a7244ee3747bb9e3792e166b9327235", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-long.26/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.07939", "identifiers": {"arxiv": "2402.07939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.07939"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-long.26/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-long.26.pdf"}], "name": "UFO: A UI-Focused Agent for Windows OS Interaction", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "UFO uses two agents to inspect Windows application interfaces and their control metadata, then grounds and executes user requests within and across applications. The study evaluates this UI-focused architecture on nine Windows applications.", "topics": ["cross-app", "planning", "grounding", "automation"], "updated_at": "2024-05-23", "url": "https://arxiv.org/abs/2402.07939"}
{"aliases": [], "authors": ["Raghav Kapoor", "Yash Parag Butala", "Melisa Russak", "Jing Yu Koh", "Kiran Kamble", "Waseem Alshikh", "Ruslan Salakhutdinov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54b548ac2bee529d3a769e7ea14b6bfce2270cdf349d832cc8c146e7d779b81a", "submitted_at": "2024-02-27T14:47:53Z", "updated_at": "2024-07-21T23:16:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.17553v3"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.17553", "identifiers": {"arxiv": "2402.17553"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.17553"}], "name": "OmniACT: A Dataset and Benchmark for Enabling Multimodal Generalist Autonomous Agents for Desktop and Web", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a dataset and benchmark for generating executable task scripts from screenshots and visually grounded desktop or web instructions.", "topics": ["evaluation", "grounding"], "updated_at": "2024-07-21", "url": "https://arxiv.org/abs/2402.17553"}
{"aliases": [], "authors": ["Shuhong Xiao", "Yunnong Chen", "Yaxuan Song", "Liuqing Chen", "Lingyun Sun", "Yankun Zhen", "Yanfang Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f4bfed18cba0c946a8ae013c819d84d7f7b2c639b2e3471d6bf7890248c4530b", "submitted_at": "2024-03-08T01:52:44Z", "updated_at": "2024-03-08T01:52:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.04984v1"}}, "environments": ["android", "ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2403.04984", "identifiers": {"arxiv": "2403.04984"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.04984"}], "name": "UI Semantic Group Detection: Grouping UI Elements with Similar Semantics in Mobile Graphical User Interface", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2024-03-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "UISCGD detects semantic groups of adjacent text and non-text UI elements. It learns from Android and iOS screenshots to support interface retrieval, UI-to-code structure and screen-reader accessibility.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-03-08", "url": "https://arxiv.org/abs/2403.04984"}
{"aliases": [], "authors": ["Longtao Zheng", "Zhiyuan Huang", "Zhenghai Xue", "Xinrun Wang", "Bo An", "Shuicheng Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe1432b46c15566182418bf3fa2f918607e8b0e9a383459416d03e28dc7e347c", "submitted_at": "2024-03-26T17:54:15Z", "updated_at": "2025-02-14T08:13:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.17918v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2403.17918", "identifiers": {"arxiv": "2403.17918"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.17918"}], "name": "AgentStudio: A Toolkit for Building General Virtual Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-03-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "AgentStudio combines an interactive GUI/API environment with tools for creating tasks, annotating interface elements, and labeling actions in videos. Its online suite and GroundUI, IDMBench, and CriticBench datasets examine interaction, grounding, learning from video, and success detection.", "topics": ["evaluation", "sandbox", "grounding", "video"], "updated_at": "2025-02-14", "url": "https://arxiv.org/abs/2403.17918"}
{"aliases": [], "authors": ["Joel Ruben Antony Moniz", "Soundarya Krishnan", "Melis Ozyildirim", "Prathamesh Saraf", "Halim Cagri Ates", "Yuan Zhang", "Hong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3573a5cf66737178f164d9809b3e6dcadb7c6b2b4ae7d95283e8a136bf78851", "submitted_at": "2024-03-29T17:59:06Z", "updated_at": "2024-08-19T03:06:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.20329v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2403.20329", "identifiers": {"arxiv": "2403.20329"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.20329"}], "name": "ReALM: Reference Resolution As Language Modeling", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. This is reference resolution involving screen entities, not an evaluated end-to-end GUI controller.", "published_at": "2024-03-29", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Frames conversational and on-screen reference resolution as language modeling, allowing textual representations of screen entities to support user-reference interpretation. Evaluation separately examines on-screen references against an existing resolver and language-model baselines.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-08-19", "url": "https://arxiv.org/abs/2403.20329"}
{"aliases": [], "authors": ["Keen You", "Haotian Zhang", "Eldon Schoop", "Floris Weers", "Amanda Swearngin", "Jeffrey Nichols", "Yinfei Yang", "Zhe Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67e139ad83ab4804dfdc4c91f3fd344d7d3d279b97efb25bd11bab1532968e46", "submitted_at": "2024-04-08T17:55:44Z", "updated_at": "2024-04-08T17:55:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05719v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.05719", "identifiers": {"arxiv": "2404.05719"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.05719"}], "name": "Ferret-UI: Grounded Mobile UI Understanding with Multimodal LLMs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts multimodal modeling to mobile interfaces through detailed screen crops, region annotations, and UI reasoning tasks.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-04-08", "url": "https://arxiv.org/abs/2404.05719"}
{"aliases": [], "authors": ["Junpeng Liu", "Yifan Song", "Bill Yuchen Lin", "Wai Lam", "Graham Neubig", "Yuanzhi Li", "Xiang Yue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7414bfa289baf86a8f2bb54f3f69f25a6a4cc8516fb0296a1899d21ad3020d4f", "submitted_at": "2024-04-09T02:29:39Z", "updated_at": "2024-04-09T02:29:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05955v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.05955", "identifiers": {"arxiv": "2404.05955"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.05955"}], "name": "VisualWebBench: How Far Have Multimodal LLMs Evolved in Web Page Understanding and Grounding?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separately evaluates web-page reading, understanding, and element grounding across tasks drawn from real websites.", "topics": ["evaluation", "screen-understanding", "grounding"], "updated_at": "2024-04-09", "url": "https://arxiv.org/abs/2404.05955"}
{"aliases": [], "authors": ["Rajat Chawla", "Adarsh Jha", "Muskaan Kumar", "Mukunda NS", "Ishaan Bhola"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e29d14a41d59942c88938b8e83217484a1cb619847c5dff0ec1f516095ee5654", "submitted_at": "2024-04-09T11:59:41Z", "updated_at": "2024-10-27T05:54:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.16048v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.16048", "identifiers": {"arxiv": "2404.16048"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.16048"}], "name": "GUIDE: Graphical User Interface Data for Execution", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces annotated website task trajectories with screenshots, prior actions, reasoning, and grounded next actions for process automation.", "topics": ["data-generation", "grounding"], "updated_at": "2024-10-27", "url": "https://arxiv.org/abs/2404.16048"}
{"aliases": [], "authors": ["Abdur Rahman", "Rajat Chawla", "Muskaan Kumar", "Arkajit Datta", "Adarsh Jha", "Mukunda NS", "Ishaan Bhola"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7cb6087a48596442e1d191bf1313e9459ad62c02c3104ee481cb72cb00e56b1", "submitted_at": "2024-05-24T08:21:45Z", "updated_at": "2024-07-21T07:34:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.15341v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:27Z", "role": "primary", "sha256": "92875dc9298b7952cc6e33ce1165b806421a6e24f3cd1e9bfd951fdd928659ad", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2405.15341v2"}], "id": "arxiv:2405.15341", "identifiers": {"arxiv": "2405.15341"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.15341"}], "name": "V-Zen: Efficient GUI Understanding and Precise Grounding With A Novel Multimodal LLM", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. The related GUIDE dataset paper remains a separate record; this annotation describes the V-Zen model contribution.", "published_at": "2024-05-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes V-Zen, a multimodal GUI model with dual-resolution image encoders for grounding and next-action prediction, fine-tuned using GUIDE interface elements and task sequences.", "topics": ["grounding", "action-prediction", "screen-understanding"], "updated_at": "2024-07-21", "url": "https://arxiv.org/abs/2405.15341"}
{"aliases": [], "authors": ["Wentong Chen", "Junbo Cui", "Jinyi Hu", "Yujia Qin", "Junjie Fang", "Yue Zhao", "Chongyi Wang", "Jun Liu", "Guirong Chen", "Yupeng Huo", "Yuan Yao", "Yankai Lin", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40e41704dbe59f1fb9b1e743a3c13394105de112213f666c2420fb235d666ef3", "submitted_at": "2024-06-17T08:30:55Z", "updated_at": "2025-05-30T07:30:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.11317v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2406.11317", "identifiers": {"arxiv": "2406.11317"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.11317"}], "name": "GUICourse: From General Vision Language Models to Versatile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUIEnv, GUIAct, and GUIChat datasets to develop OCR, grounding, component knowledge, and interaction abilities in visual-language models.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-05-30", "url": "https://arxiv.org/abs/2406.11317"}
{"aliases": [], "authors": ["Ziyang Meng", "Yu Dai", "Zezheng Gong", "Shaoxiong Guo", "Minglong Tang", "Tongquan Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea674b3aec3aee51c6f7fd4e99ad2b88394e7df43a4f3a02654d805e9a60343a", "submitted_at": "2024-06-20T07:24:43Z", "updated_at": "2024-11-04T09:31:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14056v3"}, "publication": {"authors": ["Meng Ziyang", "Yu Dai", "Zezheng Gong", "Shaoxiong Guo", "Minglong Tang", "Tongquan Wei"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.68", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.68/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "1261-1279", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.68.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "VGA: Vision GUI Assistant - Minimizing Hallucinations through Image-Centric Fine-Tuning", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.68/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "7202817c9f68a7c61afa104997fffe2a1a1112977561309ff5811f722b9411e8", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.68/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.14056", "identifiers": {"arxiv": "2406.14056"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.14056"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.68/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.68.pdf"}], "name": "VGA: Vision GUI Assistant -- Minimizing Hallucinations through Image-Centric Fine-Tuning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses image-dependent question-answer data and staged fine-tuning to improve GUI comprehension and reduce unsupported visual answers.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-11-04", "url": "https://arxiv.org/abs/2406.14056"}
{"aliases": [], "authors": ["Yue Fan", "Lei Ding", "Ching-Chen Kuo", "Shan Jiang", "Yang Zhao", "Xinze Guan", "Jie Yang", "Yi Zhang", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c86d68b5c3486b099f64792d5d3290d79e25b7c82c6eb2402869b893bb22fea", "submitted_at": "2024-06-27T15:34:16Z", "updated_at": "2024-10-25T18:16:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.19263v2"}, "publication": {"authors": ["Yue Fan", "Lei Ding", "Ching-Chen Kuo", "Shan Jiang", "Yang Zhao", "Xinze Guan", "Jie Yang", "Yi Zhang", "Xin Eric Wang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.emnlp-main.533", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.emnlp-main.533/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "9503-9522", "pdf_url": "https://aclanthology.org/2024.emnlp-main.533.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding", "type": "conference-paper", "url": "https://aclanthology.org/2024.emnlp-main.533/", "venue": "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "1d105799f5db853e801090e8ed62273cd72dd3cb9b4060f901747aace8d2406d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.emnlp-main.533/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.19263", "identifiers": {"arxiv": "2406.19263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.19263"}, {"rel": "publication", "url": "https://aclanthology.org/2024.emnlp-main.533/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.emnlp-main.533.pdf"}], "name": "Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Screen Point-and-Read and a hierarchical layout approach that interprets screen regions indicated by a user-selected point.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-10-25", "url": "https://arxiv.org/abs/2406.19263"}
{"aliases": [], "authors": ["El Hassane Ettifouri", "Jessica López Espejel", "Laura Minkova", "Tassnim Dardouri", "Walid Dahhane"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "01d3b3a013f3b10249b8197a8e18aa695bdc20febaa34f07128791f36afdcbac", "submitted_at": "2024-05-05T19:10:19Z", "updated_at": "2025-07-18T11:35:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.01558v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.01558", "identifiers": {"arxiv": "2407.01558"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.01558"}], "name": "Visual Grounding Methods for Efficient Interaction with Desktop Graphical User Interfaces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-05-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares modular OCR-based and end-to-end methods for locating desktop GUI targets from natural-language instructions.", "topics": ["grounding", "evaluation"], "updated_at": "2025-07-18", "url": "https://arxiv.org/abs/2407.01558"}
{"aliases": [], "authors": ["Yuxiang Chai", "Siyuan Huang", "Yazhe Niu", "Han Xiao", "Liang Liu", "Dingyu Zhang", "Shuai Ren", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c80665716bf665fe4688a27a4c7dcf26fcfe93cb7f79a6f28784a795e0bd3cb", "submitted_at": "2024-07-03T17:59:58Z", "updated_at": "2025-05-29T02:43:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.17490v2"}, "publication": {"authors": ["Yuxiang Chai", "Siyuan Huang", "Yazhe Niu", "Han Xiao", "Liang Liu", "Guozhi Wang", "Dingyu Zhang", "Shuai Ren", "Hongsheng Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.110", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.110/", "https://aclanthology.org/2025.findings-acl.110.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, lists nine authors including Guozhi Wang; the retained arXiv snapshot lists eight. Each edition retains its own authors.", "pages": "2138-2156", "pdf_url": "https://aclanthology.org/2025.findings-acl.110.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.110/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:52Z", "role": "primary", "sha256": "9cb5bd38791ee3df95d04815ca650bf0f020518b009c04b4a54857c122f161c8", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.110/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:02.000709+00:00", "role": "primary", "sha256": "344342e1c229e46269d26d879b32a4d2947932607b968f7e18382d231ad05ee2", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.findings-acl.110.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.17490", "identifiers": {"arxiv": "2407.17490"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.17490"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.110/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.110.pdf"}], "name": "AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Android screenshots with element locations, functionality descriptions, and instructions linked to sequences of GUI actions.", "topics": ["grounding", "screen-understanding", "mobile-navigation"], "updated_at": "2025-05-29", "url": "https://arxiv.org/abs/2407.17490"}
{"aliases": [], "authors": ["Yadong Lu", "Jianwei Yang", "Yelong Shen", "Ahmed Awadallah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfe55dd83d2f5ac48f9e3efb6bdf4aab7f11d81c0118725555456cce276adae3", "submitted_at": "2024-08-01T00:00:43Z", "updated_at": "2024-08-01T00:00:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.00203v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2408.00203", "identifiers": {"arxiv": "2408.00203"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.00203"}], "name": "OmniParser for Pure Vision Based GUI Agent", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-08-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Parses screenshots into interactable regions and semantic element descriptions to support visually grounded agent actions.", "topics": ["grounding", "screen-understanding"], "updated_at": "2024-08-01", "url": "https://arxiv.org/abs/2408.00203"}
{"aliases": [], "authors": ["Segev Shlomov", "Ben wiesel", "Aviad Sela", "Ido Levy", "Liane Galanti", "Roy Abitbol"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cdb228e652218d7f8981d9d16ca29b465004675c510d1e9c5c7b543bec9dc454", "submitted_at": "2024-09-03T14:17:09Z", "updated_at": "2024-09-03T14:17:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.01927v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2409.01927", "identifiers": {"arxiv": "2409.01927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.01927"}], "name": "From Grounding to Planning: Benchmarking Bottlenecks in Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates web-agent planning and grounding evaluations on Mind2Web to analyze which components limit task performance.", "topics": ["evaluation", "grounding", "planning"], "updated_at": "2024-09-03", "url": "https://arxiv.org/abs/2409.01927"}
{"aliases": [], "authors": ["Shuqing Li", "Binchang Li", "Yepang Liu", "Cuiyun Gao", "Jianping Zhang", "Shing-Chi Cheung", "Michael R. Lyu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a32462e4cb09763750f8f22d7f4be32a6bdfe3db6c9b3de9c386dc7cd6b6f30", "submitted_at": "2024-09-17T00:58:00Z", "updated_at": "2025-10-01T17:58:44Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2409.10811v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2409.10811", "identifiers": {"arxiv": "2409.10811"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.10811"}], "name": "Grounded GUI Understanding for Vision-Based Spatial Intelligent Agent: Exemplified by Extended Reality Apps", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced. The described environment is XR/VR applications; the operating system is not specified in the abstract.", "published_at": "2024-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Orienter detects interactable GUI elements in XR applications by interpreting scene context and revisiting candidate elements through a validation and reflection loop. It addresses open-vocabulary element types and context-dependent interactability.", "topics": ["screen-understanding", "grounding"], "updated_at": "2025-10-01", "url": "https://arxiv.org/abs/2409.10811"}
{"aliases": [], "authors": ["Wayne Chi", "Ameet Talwalkar", "Chris Donahue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2f59976cec7adb9f9c88bd79bdfbeed040e4874b6d9ab0c229136f98101aa98", "submitted_at": "2024-09-18T16:04:10Z", "updated_at": "2024-10-06T21:25:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.12089v3"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2409.12089", "identifiers": {"arxiv": "2409.12089"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.12089"}], "name": "The Impact of Element Ordering on LM Agent Performance", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2024-09-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies how the ordering of UI elements in an agent observation affects web and desktop performance. A pixel-based element detector and dimensionality-reduction ordering are evaluated on OmniACT when document hierarchy is unavailable.", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2024-10-06", "url": "https://arxiv.org/abs/2409.12089"}
{"aliases": [], "authors": ["Boyu Gou", "Ruohan Wang", "Boyuan Zheng", "Yanan Xie", "Cheng Chang", "Yiheng Shu", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08ece826d135b9ab9dfed2a3baf9deeab99489120da749cd67517b636649c5a6", "submitted_at": "2024-10-07T17:47:50Z", "updated_at": "2025-06-17T15:06:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.05243v3"}, "publication": {"authors": ["Boyu Gou", "Demi Ruohan Wang", "Boyuan Zheng", "Yanan Xie", "Cheng Chang", "Yiheng Shu", "Huan Sun", "Yu Su"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "30851-30883", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/4ca0e369689dadb25a5345ba9755ad6f-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:01Z", "role": "primary", "sha256": "4e7384659700d8bc8e1f56de671c0d943f15b8c682bdff0fbfb20f41c707b7ff", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.05243", "identifiers": {"arxiv": "2410.05243"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.05243"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/4ca0e369689dadb25a5345ba9755ad6f-Paper-Conference.pdf"}], "name": "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UGround, a visual target-localization model trained with large-scale synthetic references to elements in GUI screenshots.", "topics": ["grounding", "data-generation"], "updated_at": "2025-06-17", "url": "https://arxiv.org/abs/2410.05243"}
{"aliases": [], "authors": ["Pawel Pawlowski", "Krystian Zawistowski", "Wojciech Lapacz", "Adam Wiacek", "Marcin Skorupa", "Sebastien Postansque", "Jakub Hoscilowicz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f1594a51709f2a1283d3ac14734f08fa7894b436b85bafbbbe9b2eaba332129", "submitted_at": "2024-10-09T12:06:43Z", "updated_at": "2025-05-21T06:52:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.11871v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.11871", "identifiers": {"arxiv": "2410.11871"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.11871"}], "name": "TinyClick: Single-Turn Agent for Empowering GUI Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts a compact Florence-2 model to map user instructions to screen coordinates using multitask training and augmented data.", "topics": ["grounding", "efficiency"], "updated_at": "2025-05-21", "url": "https://arxiv.org/abs/2410.11871"}
{"aliases": [], "authors": ["Jakub Hoscilowicz", "Bartosz Maj", "Bartosz Kozakiewicz", "Oleksii Tymoshchuk", "Artur Janicki"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e804944da205076ac6cf88c7966c471697e526f657bda339e09df240257f7217", "submitted_at": "2024-10-09T14:49:02Z", "updated_at": "2024-10-17T07:12:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.11872v2"}, "publication": {"authors": ["Jakub Hoscilowicz", "Bartosz Maj", "Bartosz Kozakiewicz", "Oleksii Tymoshchuk", "Artur Janicki"], "date_basis": "Explicit publisher citation_publication_date: 2025/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 8], "doi": "10.18653/v1/2025.sigdial-1.38", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.sigdial-1.38/", "https://aclanthology.org/2025.sigdial-1.38.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the five-author list on page 1 of the publisher PDF. Landing-page citation metadata lists only Jakub Hoscilowicz and Artur Janicki.", "pages": "471-476", "pdf_url": "https://aclanthology.org/2025.sigdial-1.38.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.sigdial-1.38/", "venue": "Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:53Z", "role": "primary", "sha256": "7c710272d48754149cc5947f5382b8123eac5855db7655410259c4a3129b8b5b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.sigdial-1.38/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:09.676954+00:00", "role": "primary", "sha256": "c5b5cd07e3ca2f2911e183c4752e01c721971c47fd20c836605452f2b80e5260", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.sigdial-1.38.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.11872", "identifiers": {"arxiv": "2410.11872"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.11872"}, {"rel": "publication", "url": "https://aclanthology.org/2025.sigdial-1.38/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.sigdial-1.38.pdf"}], "name": "ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates multimodal reasoning and action planning from a dedicated UI localization model, with evaluation on Android devices.", "topics": ["mobile-navigation", "grounding"], "updated_at": "2024-10-17", "url": "https://arxiv.org/abs/2410.11872"}
{"aliases": [], "authors": ["Junpeng Liu", "Tianyue Ou", "Yifan Song", "Yuxiao Qu", "Wai Lam", "Chenyan Xiong", "Wenhu Chen", "Graham Neubig", "Xiang Yue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a3bcd26dff4f1d21a8a2772929c0c97cd1cc180a7606b0744e630dabae94609", "submitted_at": "2024-10-17T17:48:54Z", "updated_at": "2024-11-06T08:29:22Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.13824v3"}, "publication": {"authors": ["Junpeng Liu", "Tianyue Ou", "Yifan Song", "Yuxiao Qu", "Wai Lam", "Chenyan Xiong", "Wenhu Chen", "Graham Neubig", "Xiang Yue"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "17971-18005", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/2da53cd1abdae59150e35f4693834f32-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Harnessing Webpage UIs for Text-Rich Visual Understanding", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "899100e1d0cad8d949c9075f62ab3506995587b01aab3981ac4e26eaf6d88339", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.13824", "identifiers": {"arxiv": "2410.13824"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13824"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/2da53cd1abdae59150e35f4693834f32-Paper-Conference.pdf"}], "name": "Harnessing Webpage UIs for Text-Rich Visual Understanding", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-10-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds MultiUI by generating instructions from webpage accessibility trees and pairing them with screenshots for multimodal training. Evaluation examines web-interface understanding, including VisualWebBench and Mind2Web, and transfer to other text-rich visual tasks.", "topics": ["data-generation", "screen-understanding", "grounding"], "updated_at": "2024-11-06", "url": "https://arxiv.org/abs/2410.13824"}
{"aliases": [], "authors": ["Zhangheng Li", "Keen You", "Haotian Zhang", "Di Feng", "Harsh Agrawal", "Xiujun Li", "Mohana Prasad Sathya Moorthy", "Jeff Nichols", "Yinfei Yang", "Zhe Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "299b7c631468e259ac9d25a1c4ebf507be57b7ee478c611d0a2597b588f6c4c8", "submitted_at": "2024-10-24T17:58:31Z", "updated_at": "2025-02-28T00:29:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.18967v2"}}, "environments": ["ios", "android", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.18967", "identifiers": {"arxiv": "2410.18967"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.18967"}], "name": "Ferret-UI 2: Mastering Universal User Interface Understanding Across Platforms", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Extends Ferret-UI across device platforms with adaptive visual scaling and synthetic data for referring, grounding, and user-centered UI tasks.", "topics": ["screen-understanding", "grounding"], "updated_at": "2025-02-28", "url": "https://arxiv.org/abs/2410.18967"}
{"aliases": [], "authors": ["Xuetian Chen", "Hangcheng Li", "Jiaqing Liang", "Sihang Jiang", "Deqing Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dce2e22787c6e2cce95ecc4f5583bf94d7670b4b87ab75d2b8c7efb66fbcd1a8", "submitted_at": "2024-10-25T10:46:17Z", "updated_at": "2024-11-02T08:54:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.19461v2"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.19461", "identifiers": {"arxiv": "2410.19461"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.19461"}], "name": "EDGE: Enhanced Grounded GUI Understanding with Enriched Multi-Granularity Synthetic Data", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Synthesizes GUI training examples at multiple levels of detail from webpages and studies transfer of learned understanding to other interface environments.", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2024-11-02", "url": "https://arxiv.org/abs/2410.19461"}
{"aliases": [], "authors": ["Zhiyong Wu", "Zhenyu Wu", "Fangzhi Xu", "Yian Wang", "Qiushi Sun", "Chengyou Jia", "Kanzhi Cheng", "Zichen Ding", "Liheng Chen", "Paul Pu Liang", "Yu Qiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0dbd8ef8807e7886c02ed6296c9d188503fe7b676062fcff1890fd0652f10f3", "submitted_at": "2024-10-30T17:10:19Z", "updated_at": "2024-10-30T17:10:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.23218v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.23218", "identifiers": {"arxiv": "2410.23218"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.23218"}], "name": "OS-ATLAS: A Foundation Action Model for Generalist GUI Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-10-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines cross-platform GUI grounding data synthesis with action-model training to study generalization to unfamiliar interfaces.", "topics": ["grounding", "data-generation"], "updated_at": "2024-10-30", "url": "https://arxiv.org/abs/2410.23218"}
{"aliases": [], "authors": ["Xiao Liu", "Bo Qin", "Dongzhu Liang", "Guang Dong", "Hanyu Lai", "Hanchen Zhang", "Hanlin Zhao", "Iat Long Iong", "Jiadai Sun", "Jiaqi Wang", "Junjie Gao", "Junjun Shan", "Kangning Liu", "Shudan Zhang", "Shuntian Yao", "Siyi Cheng", "Wentao Yao", "Wenyi Zhao", "Xinghan Liu", "Xinyi Liu", "Xinying Chen", "Xinyue Yang", "Yang Yang", "Yifan Xu", "Yu Yang", "Yujia Wang", "Yulin Xu", "Zehan Qi", "Yuxiao Dong", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cfe1aa31818f06567aa3d619ad90ac4bdb9f9f9003a1a8a71294cd4f8ccf8dc", "submitted_at": "2024-10-28T17:05:10Z", "updated_at": "2024-10-28T17:05:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.00820v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2411.00820", "identifiers": {"arxiv": "2411.00820"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.00820"}], "name": "AutoGLM: Autonomous Foundation Agents for GUIs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents AutoGLM agents for browser and phone control, separating planning from grounding and training through an evolving interaction curriculum.", "topics": ["automation", "reinforcement-learning", "grounding"], "updated_at": "2024-10-28", "url": "https://arxiv.org/abs/2411.00820"}
{"aliases": [], "authors": ["Anthony Nguyen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b639d016daf616a5f58656d2a291505cd82f7765c47638a515e359210fd56779", "submitted_at": "2024-11-18T05:47:12Z", "updated_at": "2025-09-11T16:37:00Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2411.13591v7"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.13591", "identifiers": {"arxiv": "2411.13591"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.13591"}], "name": "Improved GUI Grounding via Iterative Narrowing", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses iterative visual narrowing to improve localization of requested GUI elements with general or specially trained vision-language models.", "topics": ["grounding"], "updated_at": "2025-09-11", "url": "https://arxiv.org/abs/2411.13591"}
{"aliases": [], "authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Zhengyuan Yang", "Shiwei Wu", "Zechen Bai", "Weixian Lei", "Lijuan Wang", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8c8f5c0a7fc5f194ffa6ca914619b9402c947d15f100b124505883ad8ff23ac7", "submitted_at": "2024-11-26T14:29:47Z", "updated_at": "2024-11-26T14:29:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.17465v1"}, "publication": {"authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Zhengyuan Yang", "Shiwei Wu", "Zechen Bai", "Stan Weixian Lei", "Lijuan Wang", "Mike Zheng Shou"], "date_basis": "Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.", "date_parts": [2025, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher page directly links this arXiv identifier.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "19498-19508", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2025/papers/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.pdf", "publisher": null, "series": null, "title": "ShowUI: One Vision-Language-Action Model for GUI Visual Agent", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:00Z", "role": "primary", "sha256": "2a79195c42f55ef363988febb2b7cae1d348c4450a443462b75c96fd20c17256", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2411.17465", "identifiers": {"arxiv": "2411.17465"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.17465"}, {"rel": "publication", "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"}, {"rel": "publication_pdf", "url": "https://openaccess.thecvf.com/content/CVPR2025/papers/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.pdf"}], "name": "ShowUI: One Vision-Language-Action Model for GUI Visual Agent", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-11-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines visual token selection, interleaved visual-action history, and curated instruction data in a compact GUI action model.", "topics": ["grounding", "efficiency"], "updated_at": "2024-11-26", "url": "https://arxiv.org/abs/2411.17465"}
{"aliases": [], "authors": ["Yiqin Wang", "Haoji Zhang", "Jingqi Tian", "Yansong Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a4c750017de421d3ca9bf119834dfb96d2fe8a0610f12b30d2fae603e26c66b", "submitted_at": "2024-12-02T08:35:31Z", "updated_at": "2024-12-02T08:35:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.01268v1"}, "publication": {"authors": ["Yiqin Wang", "Haoji Zhang", "Jingqi Tian", "Yansong Tang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.76", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.76/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "1461-1473", "pdf_url": "https://aclanthology.org/2025.findings-acl.76.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Ponder & Press: Advancing Visual GUI Agent towards General Computer Control", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.76/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["cross-platform", "web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:57Z", "role": "primary", "sha256": "0ec2dd7116411e80f842627d047d06cf60d5921d91350daeb251a9c95a8f7884", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.76/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.01268", "identifiers": {"arxiv": "2412.01268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.01268"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.76/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.76.pdf"}], "name": "Ponder & Press: Advancing Visual GUI Agent towards General Computer Control", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates interpretation of user instructions from precise GUI localization using general-purpose and specialized visual-language components.", "topics": ["grounding", "planning"], "updated_at": "2024-12-02", "url": "https://arxiv.org/abs/2412.01268"}
{"aliases": [], "authors": ["Yiheng Xu", "Zekun Wang", "Junli Wang", "Dunjie Lu", "Tianbao Xie", "Amrita Saha", "Doyen Sahoo", "Tao Yu", "Caiming Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89dbdf874ace87db12fa321e67ecf818d3964622b045ea7d95e109101820d407", "submitted_at": "2024-12-05T18:58:26Z", "updated_at": "2025-05-05T16:17:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.04454v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.04454", "identifiers": {"arxiv": "2412.04454"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.04454"}], "name": "Aguvis: Unified Pure Vision Agents for Autonomous GUI Interaction", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-12-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses screenshots, a unified action space, and staged grounding and reasoning training for cross-platform GUI interaction.", "topics": ["grounding", "planning", "data-generation"], "updated_at": "2025-05-05", "url": "https://arxiv.org/abs/2412.04454"}
{"aliases": [], "authors": ["Zhiqi Ge", "Juncheng Li", "Xinglei Pang", "Minghe Gao", "Kaihang Pan", "Wang Lin", "Hao Fei", "Wenqiao Zhang", "Siliang Tang", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e7cab57b118ab76a9026e7ee800709f4e93969e19ad18f549f8b75ee657622b", "submitted_at": "2024-12-13T18:40:10Z", "updated_at": "2025-02-03T15:23:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.10342v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.10342", "identifiers": {"arxiv": "2412.10342"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10342"}], "name": "Iris: Breaking GUI Complexity with Adaptive Focus and Self-Refining", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines information-sensitive image cropping with a learning loop between element description and grounding for complex interface understanding.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2025-02-03", "url": "https://arxiv.org/abs/2412.10342"}
{"aliases": [], "authors": ["Jiarun Liu", "Jia Hao", "Chunhong Zhang", "Zheng Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2abbae042d20381aa218e6a2b14a1fc28074543e430ad09031237d22587be7a", "submitted_at": "2024-12-14T08:25:28Z", "updated_at": "2024-12-14T08:25:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10742v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.10742", "identifiers": {"arxiv": "2412.10742"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10742"}], "name": "WEPO: Web Element Preference Optimization for LLM-based Web Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses web-element preference optimization with sampled negative HTML elements to align high-level user intent with navigation actions.", "topics": ["grounding", "web-navigation"], "updated_at": "2024-12-14", "url": "https://arxiv.org/abs/2412.10742"}
{"aliases": [], "authors": ["Hai-Ming Xu", "Qi Chen", "Lei Wang", "Lingqiao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6e0aefcd3eca9e37d25a2c5195c0d502caec9be83b8f6b4b7ee5778bff9e150", "submitted_at": "2024-12-14T14:30:05Z", "updated_at": "2024-12-14T14:30:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10840v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.10840", "identifiers": {"arxiv": "2412.10840"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10840"}], "name": "Attention-driven GUI Grounding: Leveraging Pretrained Multimodal Large Language Models without Fine-Tuning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Locates GUI targets by identifying and aggregating attention maps in pretrained multimodal models without task-specific fine-tuning.", "topics": ["grounding"], "updated_at": "2024-12-14", "url": "https://arxiv.org/abs/2412.10840"}
{"aliases": [], "authors": ["Yuhao Yang", "Yue Wang", "Dongxu Li", "Ziyang Luo", "Bei Chen", "Chao Huang", "Junnan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b49d45aa539c224846a7e14cee9b537d032228e19e86a9cb3cdd9f5004aa429", "submitted_at": "2024-12-20T07:16:57Z", "updated_at": "2025-07-08T08:49:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.16256v2"}, "publication": {"authors": ["Yuhao Yang", "Yue Wang", "Dongxu Li", "Ziyang Luo", "Bei Chen", "Chao Huang", "Junnan Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.1152", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.1152/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "22418-22433", "pdf_url": "https://aclanthology.org/2025.findings-acl.1152.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Aria-UI: Visual Grounding for GUI Instructions", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.1152/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:57Z", "role": "primary", "sha256": "e24a63769d700caa2d27341dd2634b990cdbc53c7cf557817b3679dc63faef6f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.1152/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.16256", "identifiers": {"arxiv": "2412.16256"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.16256"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.1152/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.1152.pdf"}], "name": "Aria-UI: Visual Grounding for GUI Instructions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a vision-based grounding model trained with diverse synthetic instructions and action histories for context-sensitive GUI targeting.", "topics": ["grounding", "memory"], "updated_at": "2025-07-08", "url": "https://arxiv.org/abs/2412.16256"}
{"aliases": [], "authors": ["Yuhang Liu", "Pengxiang Li", "Zishu Wei", "Congkai Xie", "Xueyu Hu", "Xinchen Xu", "Shengyu Zhang", "Xiaotian Han", "Hongxia Yang", "Fei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "18d5b2e34f18c24e5d60d5564c9b08eced39c1b17bf5dac8d997640786688fba", "submitted_at": "2025-01-08T15:45:21Z", "updated_at": "2025-01-08T15:45:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.04575v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2501.04575", "identifiers": {"arxiv": "2501.04575"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.04575"}], "name": "InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses staged supervised training to combine GUI understanding and grounding with hierarchical reasoning and expectation-based reflection.", "topics": ["grounding", "planning"], "updated_at": "2025-01-08", "url": "https://arxiv.org/abs/2501.04575"}
{"aliases": [], "authors": ["Yujia Qin", "Yining Ye", "Junjie Fang", "Haoming Wang", "Shihao Liang", "Shizuo Tian", "Junda Zhang", "Jiahao Li", "Yunxin Li", "Shijue Huang", "Wanjun Zhong", "Kuanye Li", "Jiale Yang", "Yu Miao", "Woyu Lin", "Longxiang Liu", "Xu Jiang", "Qianli Ma", "Jingyu Li", "Xiaojun Xiao", "Kai Cai", "Chuang Li", "Yaowei Zheng", "Chaolin Jin", "Chen Li", "Xiao Zhou", "Minchao Wang", "Haoli Chen", "Zhaojian Li", "Haihua Yang", "Haifeng Liu", "Feng Lin", "Tao Peng", "Xin Liu", "Guang Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d19aa758e66411a31dd1fea6f07c83d1c828bd4aad86f1564a9d7f3670c3f79f", "submitted_at": "2025-01-21T17:48:10Z", "updated_at": "2025-01-21T17:48:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.12326v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2501.12326", "identifiers": {"arxiv": "2501.12326"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.12326"}], "name": "UI-TARS: Pioneering Automated GUI Interaction with Native Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-01-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies a screenshot-driven native GUI model with unified actions, deliberate reasoning, and iterative training on interaction traces.", "topics": ["grounding", "planning", "data-generation"], "updated_at": "2025-01-21", "url": "https://arxiv.org/abs/2501.12326"}
{"aliases": [], "authors": ["Yue Fan", "Handong Zhao", "Ruiyi Zhang", "Yu Shen", "Xin Eric Wang", "Gang Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0804055deb6c3513990ed8d596486400f1cb791338b7866d9e949186a34eb9c5", "submitted_at": "2025-01-23T18:16:21Z", "updated_at": "2025-01-27T18:58:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.13896v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2501.13896", "identifiers": {"arxiv": "2501.13896"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.13896"}], "name": "GUI-Bee: Align GUI Action Grounding to Novel Environments via Autonomous Exploration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Collects environment-specific interaction data through autonomous exploration to adapt grounding models to previously unseen GUIs.", "topics": ["grounding", "data-generation"], "updated_at": "2025-01-27", "url": "https://arxiv.org/abs/2501.13896"}
{"aliases": [], "authors": ["Hongxin Li", "Jingfan Chen", "Jingran Su", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b76ab1c6c44db9f86b2bfc11097c54e5c3115c0c498a58d896bef043d2113c8", "submitted_at": "2025-02-04T03:39:59Z", "updated_at": "2025-06-07T05:23:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.01977v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.01977", "identifiers": {"arxiv": "2502.01977"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.01977"}], "name": "AutoGUI: Scaling GUI Grounding with Automatic Functionality Annotations from LLMs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Infers element functionality from UI changes before and after interaction to produce verified annotations for grounding-model training.", "topics": ["grounding", "data-generation"], "updated_at": "2025-06-07", "url": "https://arxiv.org/abs/2502.01977"}
{"aliases": [], "authors": ["Kunal Singh", "Shreyas Singh", "Mukund Khanna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a60d581e4318db66495a0333b864757ec896576c720e40da4a881f8740b9d9cf", "submitted_at": "2025-02-12T09:12:30Z", "updated_at": "2025-02-14T06:23:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.08226v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.08226", "identifiers": {"arxiv": "2502.08226"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.08226"}], "name": "TRISHUL: Towards Region Identification and Screen Hierarchy Understanding for Large VLM based GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines hierarchical screen parsing with spatially enriched element descriptions to support both locating GUI targets and describing indicated elements.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-02-14", "url": "https://arxiv.org/abs/2502.08226"}
{"aliases": [], "authors": ["Zongru Wu", "Pengzhou Cheng", "Zheng Wu", "Tianjie Ju", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53781edbb107437b258d0c0e298f80ec14f8f32b51bc0dd11c0417473eb2c606", "submitted_at": "2025-03-01T08:29:59Z", "updated_at": "2025-03-04T12:04:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.00401v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.00401", "identifiers": {"arxiv": "2503.00401"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.00401"}], "name": "Smoothing Grounding and Reasoning for MLLM-Powered GUI Agents with Query-Oriented Pivot Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses inference of possible user queries from screenshots and target coordinates as a training bridge between GUI grounding and action reasoning.", "topics": ["grounding", "planning"], "updated_at": "2025-03-04", "url": "https://arxiv.org/abs/2503.00401"}
{"aliases": [], "authors": ["Zhiyuan Huang", "Ziming Cheng", "Junting Pan", "Zhaohui Hou", "Mingjie Zhan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d45cacb33df18242eb3f47f6c89c56e41a146dd110e4b4dc5340062b9b981250", "submitted_at": "2025-03-05T05:30:22Z", "updated_at": "2025-04-16T15:34:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.03196v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.03196", "identifiers": {"arxiv": "2503.03196"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.03196"}], "name": "SpiritSight Agent: Advanced GUI Agent with One Look", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SpiritSight and GUI-Lasagne data, combining multilevel GUI annotations with visual block parsing for high-resolution interface navigation.", "topics": ["grounding", "screen-understanding", "data-generation"], "updated_at": "2025-04-16", "url": "https://arxiv.org/abs/2503.03196"}
{"aliases": [], "authors": ["Zheng Hui", "Yinheng Li", "Dan zhao", "Tianyi Chen", "Colby Banbury", "Kazuhito Koishida"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc499e361f0bc46c3c5ee5a8abebd911396e383a004655de56db1ec7cec8e231", "submitted_at": "2025-01-27T08:29:17Z", "updated_at": "2025-01-27T08:29:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.04730v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.04730", "identifiers": {"arxiv": "2503.04730"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.04730"}], "name": "WinClick: GUI Grounding with Multimodal Large Language Models", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a screenshot-based Windows GUI agent and the WinSpot benchmark for evaluating localization of actionable interface regions.", "topics": ["grounding", "evaluation"], "updated_at": "2025-01-27", "url": "https://arxiv.org/abs/2503.04730"}
{"aliases": [], "authors": ["Fei Tang", "Yongliang Shen", "Hang Zhang", "Siqi Chen", "Guiyang Hou", "Wenqi Zhang", "Wenqiao Zhang", "Kaitao Song", "Weiming Lu", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebc104ee6a868092ae0215f84a3c5c71894bfb321671476a74b794e1645210c7", "submitted_at": "2025-03-09T06:14:17Z", "updated_at": "2025-03-09T06:14:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.06470v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.06470", "identifiers": {"arxiv": "2503.06470"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.06470"}], "name": "Think Twice, Click Once: Enhancing GUI Grounding via Fast and Slow Systems", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines rapid target prediction with deliberate interface analysis, adaptively selecting the grounding process according to task complexity.", "topics": ["grounding", "efficiency", "planning"], "updated_at": "2025-03-09", "url": "https://arxiv.org/abs/2503.06470"}
{"aliases": [], "authors": ["Yibin Xu", "Liang Yang", "Hao Chen", "Hua Wang", "Zhi Chen", "Yaohua Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2e6250e87ea9340914b457c4b0d471450f8cf49d2e2d53abd8e9e76105e908a", "submitted_at": "2025-03-14T08:16:02Z", "updated_at": "2025-03-14T08:16:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.11170v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.11170", "identifiers": {"arxiv": "2503.11170"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.11170"}], "name": "DeskVision: Large Scale Desktop Region Captioning for Advanced GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AutoCaptioner, DeskVision, and GUIExplorer for generating desktop region descriptions and training interface-understanding models.", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2025-03-14", "url": "https://arxiv.org/abs/2503.11170"}
{"aliases": [], "authors": ["Ziwei Wang", "Weizhi Chen", "Leyang Yang", "Sheng Zhou", "Shengchu Zhao", "Hanbei Zhan", "Jiongchao Jin", "Liangcheng Li", "Zirui Shao", "Jiajun Bu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b260fd1d7e28c8263710f116638e7aa7b51f46aac15199c8a29092161a13f37d", "submitted_at": "2025-03-18T08:32:22Z", "updated_at": "2025-03-18T08:32:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.14021v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2503.14021", "identifiers": {"arxiv": "2503.14021"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.14021"}], "name": "MP-GUI: Modality Perception with MLLMs for GUI Understanding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses specialized graphical, textual, and spatial perception components with a data-collection pipeline to improve GUI understanding.", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2025-03-18", "url": "https://arxiv.org/abs/2503.14021"}
{"aliases": [], "authors": ["Shravan Nayak", "Xiangru Jian", "Kevin Qinghong Lin", "Juan A. Rodriguez", "Montek Kalsi", "Rabiul Awal", "Nicolas Chapados", "M. Tamer Özsu", "Aishwarya Agrawal", "David Vazquez", "Christopher Pal", "Perouz Taslakian", "Spandana Gella", "Sai Rajeswar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5abf3c7b08a4331d4f634d8cc63a7f8ff40713f7344231887dac1b5e061a8eae", "submitted_at": "2025-03-19T19:26:17Z", "updated_at": "2025-05-06T17:43:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.15661v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2503.15661", "identifiers": {"arxiv": "2503.15661"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.15661"}], "name": "UI-Vision: A Desktop-centric GUI Benchmark for Visual Perception and Interaction", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-03-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Provides annotated desktop demonstrations for evaluating element grounding, layout grounding, and action prediction.", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2025-05-06", "url": "https://arxiv.org/abs/2503.15661"}
{"aliases": [], "authors": ["Zhengxi Lu", "Yuxiang Chai", "Yaxuan Guo", "Xi Yin", "Liang Liu", "Hao Wang", "Han Xiao", "Shuai Ren", "Guanjing Xiong", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b794a2f59cbe26a889b242ff47a7cd982f6367f759ddabcf99d96f31e5c122da", "submitted_at": "2025-03-27T15:39:30Z", "updated_at": "2025-05-24T08:46:08Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2503.21620v5"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.21620", "identifiers": {"arxiv": "2503.21620"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.21620"}], "name": "UI-R1: Enhancing Efficient Action Prediction of GUI Agents by Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses rule-based action rewards and policy optimization to train GUI action prediction from a compact set of mobile tasks.", "topics": ["reinforcement-learning", "grounding"], "updated_at": "2025-05-24", "url": "https://arxiv.org/abs/2503.21620"}
{"aliases": [], "authors": ["Saaket Agashe", "Kyle Wong", "Vincent Tu", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "11958f69503db514a91af8721870a4cbc1fa19dabf94e3f4ce4118d2fad9862a", "submitted_at": "2025-04-01T15:40:27Z", "updated_at": "2025-04-01T15:40:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.00906v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.00906", "identifiers": {"arxiv": "2504.00906"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.00906"}], "name": "Agent S2: A Compositional Generalist-Specialist Framework for Computer Use Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-04-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines generalist and specialist models with hierarchical planning and multiple grounding methods for computer-use tasks.", "topics": ["grounding", "planning"], "updated_at": "2025-04-01", "url": "https://arxiv.org/abs/2504.00906"}
{"aliases": [], "authors": ["Haoren Zhao", "Tianyi Chen", "Zhen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ededfa864150f1b7b4ba53b265b5cbaa60bcf42690a8e22f5c13f915b6cd0108", "submitted_at": "2025-04-07T03:58:45Z", "updated_at": "2025-04-07T03:58:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.04716v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.04716", "identifiers": {"arxiv": "2504.04716"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.04716"}], "name": "On the Robustness of GUI Grounding Models Against Image Attacks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates GUI grounding models under natural image noise, untargeted adversarial perturbations, and targeted image attacks. Experiments cover mobile, desktop, and web interfaces and analyze sensitivity to image resolution and adversarial changes.", "topics": ["grounding", "robustness", "safety-security", "evaluation"], "updated_at": "2025-04-07", "url": "https://arxiv.org/abs/2504.04716"}
{"aliases": [], "authors": ["Kaixin Li", "Ziyang Meng", "Hongzhan Lin", "Ziyang Luo", "Yuchen Tian", "Jing Ma", "Zhiyong Huang", "Tat-Seng Chua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "679bfcac736544e9a241c579adfda0cacbc40838452ea56ad631f8086868415a", "submitted_at": "2025-04-04T14:25:17Z", "updated_at": "2025-04-04T14:25:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.07981v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.07981", "identifiers": {"arxiv": "2504.07981"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.07981"}], "name": "ScreenSpot-Pro: GUI Grounding for Professional High-Resolution Computer Use", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-04-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates GUI target localization in high-resolution professional software and studies cascaded visual search to narrow the target region.", "topics": ["grounding", "evaluation"], "updated_at": "2025-04-04", "url": "https://arxiv.org/abs/2504.07981"}
{"aliases": [], "authors": ["Iason Chaimalas", "Arnas Vyšniauskas", "Gabriel Brostow"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf23700bf7626733b0016b773852455c503617adc81025fc9e097879b5534cca", "submitted_at": "2025-04-12T22:02:29Z", "updated_at": "2025-04-12T22:02:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.09352v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2504.09352", "identifiers": {"arxiv": "2504.09352"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.09352"}], "name": "Explorer: Robust Collection of Interactable GUI Elements", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Explorer collects interactable GUI elements from a live target application and records user sessions for application-specific model training. It links repeated interface states into a navigation map that supports path planning and voice-directed interaction.", "topics": ["data-generation", "recording", "grounding", "planning"], "updated_at": "2025-04-12", "url": "https://arxiv.org/abs/2504.09352"}
{"aliases": [], "authors": ["Suyu Ye", "Haojun Shi", "Darren Shih", "Hyokun Yun", "Tanya Roosta", "Tianmin Shu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "268adf347e1f224f434e6dba838873c99a4050686cae1495643d9638d840d809", "submitted_at": "2025-04-14T17:36:46Z", "updated_at": "2025-12-01T15:22:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.10445v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2504.10445", "identifiers": {"arxiv": "2504.10445"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.10445"}], "name": "RealWebAssist: A Benchmark for Long-Horizon Web Assistance with Real-World Users", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "RealWebAssist evaluates web assistance through sequences of instructions collected from real users across multiple websites. It tests intent interpretation, user-specific routines, long-term context, and visual grounding when instructions are ambiguous or evolve during interaction.", "topics": ["evaluation", "long-horizon", "human-in-the-loop", "grounding"], "updated_at": "2025-12-01", "url": "https://arxiv.org/abs/2504.10445"}
{"aliases": [], "authors": ["Run Luo", "Lu Wang", "Wanwei He", "Longze Chen", "Jiaming Li", "Xiaobo Xia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b875617b795e0f8d7d6348e22ec85ccc9ab4dd931e66850182348927b1e5209f", "submitted_at": "2025-04-14T17:45:54Z", "updated_at": "2025-10-01T04:55:20Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.10458v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.10458", "identifiers": {"arxiv": "2504.10458"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.10458"}], "name": "GUI-R1 : A Generalist R1-Style Vision-Language Action Model For GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Applies reinforcement fine-tuning with unified action rules and curated cross-platform examples to vision-language GUI agents.", "topics": ["reinforcement-learning", "grounding"], "updated_at": "2025-10-01", "url": "https://arxiv.org/abs/2504.10458"}
{"aliases": [], "authors": ["Xinyi Liu", "Xiaoyi Zhang", "Ziyun Zhang", "Yan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "363e3c3e882fa93a19df74bbf77aa98525cf4639c8f62100c3ddc3e0a438997f", "submitted_at": "2025-04-15T14:56:21Z", "updated_at": "2025-07-30T13:57:49Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.11257v4"}, "publication": {"authors": ["Xinyi Liu", "Xiaoyi Zhang", "Ziyun Zhang", "Yan Lu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.809", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.809/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "15668-15684", "pdf_url": "https://aclanthology.org/2025.findings-acl.809.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.809/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:59Z", "role": "primary", "sha256": "f7bcea2d3656ebc3ac305d65271000e411b735eb10556d23ee0d6157593a109f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.809/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.11257", "identifiers": {"arxiv": "2504.11257"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11257"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.809/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.809.pdf"}], "name": "UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-E2I-Synth uses GPT-4o to synthesize GUI grounding instructions with varied complexity, addressing small elements, imbalanced element types, and implicit requests. The accompanying UI-I2E-Bench evaluates instruction-to-element localization across these annotation challenges.", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-07-30", "url": "https://arxiv.org/abs/2504.11257"}
{"aliases": [], "authors": ["Bofei Zhang", "Zirui Shang", "Zhi Gao", "Wang Zhang", "Rui Xie", "Xiaojian Ma", "Tao Yuan", "Xinxiao Wu", "Song-Chun Zhu", "Qing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "346fc4db7f8fbf2dc66c32e36f5cc5ace22228d60bdd4c61b6b7ac65487a4a8c", "submitted_at": "2025-04-17T06:15:56Z", "updated_at": "2025-12-04T15:01:09Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.12679v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.12679", "identifiers": {"arxiv": "2504.12679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.12679"}], "name": "TongUI: Internet-Scale Trajectories from Multimodal Web Tutorials for Generalized GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "TongUI converts multimodal online tutorials, including videos and articles, into GUI-agent trajectories. Its GUI-Net dataset contains 143,000 trajectories spanning five operating systems and more than 200 applications, and supports fine-tuning for grounding and navigation.", "topics": ["data-generation", "grounding", "video"], "updated_at": "2025-12-04", "url": "https://arxiv.org/abs/2504.12679"}
{"aliases": [], "authors": ["Zhiyuan Hu", "Shiyun Xiong", "Yifan Zhang", "See-Kiong Ng", "Anh Tuan Luu", "Bo An", "Shuicheng Yan", "Bryan Hooi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1353bb4651f449763ac58803b4858edf6e1d5b0ad3ab3e4c333883a1ebbaebf4", "submitted_at": "2025-04-22T17:52:42Z", "updated_at": "2025-04-22T17:52:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.16073v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:10Z", "role": "primary", "sha256": "b1cf161fc20efc51363d9007952ce5f917963d4ab2ca32d39099229060b29528", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2504.16073v1"}], "id": "arxiv:2504.16073", "identifiers": {"arxiv": "2504.16073"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.16073"}], "name": "Guiding VLM Agents with Process Rewards at Inference Time for GUI Navigation", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-04-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Guides visual GUI agents at inference time with a process-reward model that evaluates intermediate actions. It combines stepwise guidance with trajectory reflection and retry mechanisms in static and interactive navigation tasks.", "topics": ["reward-models", "self-reflection", "grounding"], "updated_at": "2025-04-22", "url": "https://arxiv.org/abs/2504.16073"}
{"aliases": [], "authors": ["Qi Yang", "Weichen Bi", "Haiyang Shen", "Yaoqi Guo", "Yun Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1ec3311094166baac4de66ad0f083b66585c933f19cb66fc580bf128ad809ca", "submitted_at": "2025-04-23T05:01:25Z", "updated_at": "2025-04-27T06:52:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.16419v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:13Z", "role": "primary", "sha256": "496c0419e2c8c752fc2c88d015ad2cf13c182aa4501a8305ff65ad4178113843", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2504.16419v2"}], "id": "arxiv:2504.16419", "identifiers": {"arxiv": "2504.16419"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.16419"}], "name": "PixelWeb: The First Web GUI Dataset with Pixel-Wise Labels", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-04-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PixelWeb, a webpage GUI dataset with element images, contours, and pixel masks. Its annotation pipeline combines bitmap channel information with DOM-based visibility and stacking analysis to address occlusion and overlapping elements.", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2025-04-27", "url": "https://arxiv.org/abs/2504.16419"}
{"aliases": [], "authors": ["Jing Huang", "Zhixiong Zeng", "Wenkang Han", "Yufeng Zhong", "Liming Zheng", "Shuai Fu", "Jingyuan Chen", "Lin Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e582c92521ec41a0a3daa0ca72b8c00250621289352b28400ee6d311c4a106eb", "submitted_at": "2025-05-01T09:27:13Z", "updated_at": "2025-05-01T09:27:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.00416v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.00416", "identifiers": {"arxiv": "2505.00416"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.00416"}], "name": "ScaleTrack: Scaling and back-tracking Automated GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScaleTrack combines a unified collection of GUI grounding samples with a planning objective that predicts the next action while reconstructing prior actions. This backward prediction links the current screenshot to interaction history during agent training.", "topics": ["grounding", "planning", "data-generation", "interaction-history"], "updated_at": "2025-05-01", "url": "https://arxiv.org/abs/2505.00416"}
{"aliases": [], "authors": ["Tiange Luo", "Lajanugen Logeswaran", "Justin Johnson", "Honglak Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d2dc8ca397adce1ecf6cf53afa82a8d4f1825bf00277aeb117966f965abfe48", "submitted_at": "2025-05-01T17:45:59Z", "updated_at": "2025-07-14T17:45:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.00684v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.00684", "identifiers": {"arxiv": "2505.00684"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.00684"}], "name": "Visual Test-time Scaling for GUI Agent Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "RegionFocus adds visual test-time scaling by zooming into selected screenshot regions and maintaining an image-based map of landmarks and actions. It evaluates this grounding procedure with UI-TARS and Qwen2.5-VL on ScreenSpot-Pro and WebVoyager.", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2025-07-14", "url": "https://arxiv.org/abs/2505.00684"}
{"aliases": [], "authors": ["Xinbin Yuan", "Jian Zhang", "Kaixin Li", "Zhuoxuan Cai", "Lujian Yao", "Jie Chen", "Enguang Wang", "Qibin Hou", "Jinwei Chen", "Peng-Tao Jiang", "Bo Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755f885b79e9f3eef9b4cbdb8593fb3cde1ac8a92f86df52d5d4203eb6cad039", "submitted_at": "2025-05-18T11:22:04Z", "updated_at": "2025-05-24T03:01:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.12370v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.12370", "identifiers": {"arxiv": "2505.12370"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12370"}], "name": "Enhancing Visual Grounding for GUI Agents via Self-Evolutionary Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes GUI grounding training that combines curated seed examples, continuous location-dependent policy feedback, and iterative refinement guided by attention maps. The study evaluates the resulting reinforcement-learning procedure on three grounding benchmarks, including ScreenSpot-Pro.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-24", "url": "https://arxiv.org/abs/2505.12370"}
{"aliases": [], "authors": ["Longxi Gao", "Li Zhang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Mengwei Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10d826ea7715ab7054522a9dcba622d5dae3e6c3faee421e4b168e16104d184b", "submitted_at": "2025-05-18T16:34:30Z", "updated_at": "2025-10-10T09:28:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.12493v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.12493", "identifiers": {"arxiv": "2505.12493"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12493"}], "name": "GUI-Shift: Enhancing VLM-Based GUI Agents through Self-supervised Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Shift trains an inverse-dynamics task that predicts the initial action connecting two GUI states, using trajectories without natural-language instructions. It combines rule-based reinforcement learning with data filtering and evaluates transfer to Android automation and screenshot grounding.", "topics": ["reinforcement-learning", "data-generation", "grounding", "mobile-navigation"], "updated_at": "2025-10-10", "url": "https://arxiv.org/abs/2505.12493"}
{"aliases": [], "authors": ["Tianbao Xie", "Jiaqi Deng", "Xiaochuan Li", "Junlin Yang", "Haoyuan Wu", "Jixuan Chen", "Wenjing Hu", "Xinyuan Wang", "Yuhui Xu", "Zekun Wang", "Yiheng Xu", "Junli Wang", "Doyen Sahoo", "Tao Yu", "Caiming Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bec4ad578c01ebdfb58bd269562d3ca7a2bd8d2687cde07cb867db0b78cb105", "submitted_at": "2025-05-19T15:09:23Z", "updated_at": "2025-10-24T19:08:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.13227v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.13227", "identifiers": {"arxiv": "2505.13227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.13227"}], "name": "Scaling Computer-Use Grounding via User Interface Decomposition and Synthesis", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OSWorld-G grounding evaluation and the Jedi synthesis pipeline for examples spanning layout reasoning and fine-grained interaction.", "topics": ["grounding", "data-generation", "evaluation"], "updated_at": "2025-10-24", "url": "https://arxiv.org/abs/2505.13227"}
{"aliases": [], "authors": ["Hyunseok Lee", "Jeonghoon Kim", "Beomjun Kim", "Jihoon Tack", "Chansong Jo", "Jaehong Lee", "Cheonbok Park", "Sookyo In", "Jinwoo Shin", "Kang Min Yoo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9acb7a88a0cf07427c1878c896dec920852810879c30bad8692358f2e2016539", "submitted_at": "2025-05-21T08:36:18Z", "updated_at": "2025-05-24T15:08:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15259v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.15259", "identifiers": {"arxiv": "2505.15259"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.15259"}], "name": "ReGUIDE: Data Efficient GUI Grounding via Spatial Reasoning and Search", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines self-generated localization reasoning, spatial consistency checks, and inference-time search to improve data-efficient web grounding.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-24", "url": "https://arxiv.org/abs/2505.15259"}
{"aliases": [], "authors": ["Yuqi Zhou", "Sunhao Dai", "Shuai Wang", "Kaiwen Zhou", "Qinglin Jia", "Jun Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "321f2d627451d60e130c17574994648fae69fc784b0ed03f6560e7f1ef9eb023", "submitted_at": "2025-05-21T17:59:09Z", "updated_at": "2025-05-22T11:15:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15810v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.15810", "identifiers": {"arxiv": "2505.15810"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.15810"}], "name": "GUI-G1: Understanding R1-Zero-Like Training for Visual Grounding in GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes reasoning templates, reward design, and policy updates in GUI grounding, then proposes targeted adjustments to the training process.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-22", "url": "https://arxiv.org/abs/2505.15810"}
{"aliases": [], "authors": ["Shaoqing Zhang", "Kehai Chen", "Zhuosheng Zhang", "Rumei Li", "Rongxiang Weng", "Yang Xiang", "Min Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4fe40099905712f9d7ba89bc7616d09b260fe6113beb8122cade54a66d46e3d7", "submitted_at": "2025-05-27T14:49:30Z", "updated_at": "2025-09-26T09:13:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.21279v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2505.21279", "identifiers": {"arxiv": "2505.21279"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21279"}], "name": "XBOUND: Exploring Capability Boundaries of Device-Control Agents at the State Level", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "XBOUND evaluates GUI agents at the level of a screen state, considering multiple instructions and possible target actions within it. The study separates action matching from instruction unification and compares effects of grounding and trajectory data.", "topics": ["evaluation", "grounding"], "updated_at": "2025-09-26", "url": "https://arxiv.org/abs/2505.21279"}
{"aliases": [], "authors": ["Mathieu Andreux", "Breno Baldas Skuk", "Hamza Benchekroun", "Emilien Biré", "Antoine Bonnet", "Riaz Bordie", "Nathan Bout", "Matthias Brunel", "Pierre-Louis Cedoz", "Antoine Chassang", "Mickaël Chen", "Alexandra D. Constantinou", "Antoine d'Andigné", "Hubert de La Jonquière", "Aurélien Delfosse", "Ludovic Denoyer", "Alexis Deprez", "Augustin Derupti", "Michael Eickenberg", "Mathïs Federico", "Charles Kantor", "Xavier Koegler", "Yann Labbé", "Matthew C. H. Lee", "Erwan Le Jumeau de Kergaradec", "Amir Mahla", "Avshalom Manevich", "Adrien Maret", "Charles Masson", "Rafaël Maurin", "Arturo Mena", "Philippe Modard", "Axel Moyal", "Axel Nguyen Kerbel", "Julien Revelle", "Mats L. Richter", "María Santos", "Laurent Sifre", "Maxime Theillard", "Marc Thibault", "Louis Thiry", "Léo Tronchon", "Nicolas Usunier", "Tony Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89c2d658e03204b56f6d273da9572820f9a23ca359a441ea87baa451a93c5b58", "submitted_at": "2025-06-03T13:29:03Z", "updated_at": "2025-06-11T09:33:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.02865v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.02865", "identifiers": {"arxiv": "2506.02865"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02865"}], "name": "Surfer-H Meets Holo1: Cost-Efficient Web Agent Powered by Open Weights", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes Surfer-H, a web agent paired with the Holo1 vision-language model family trained on web content, synthetic examples, and agent-generated data. The work introduces WebClick for web-interface localization and evaluates browser task execution on WebVoyager.", "topics": ["web-navigation", "grounding", "data-generation", "evaluation"], "updated_at": "2025-06-11", "url": "https://arxiv.org/abs/2506.02865"}
{"aliases": [], "authors": ["Qianhui Wu", "Kanzhi Cheng", "Rui Yang", "Chaoyun Zhang", "Jianwei Yang", "Huiqiang Jiang", "Jian Mu", "Baolin Peng", "Bo Qiao", "Reuben Tan", "Si Qin", "Lars Liden", "Qingwei Lin", "Huan Zhang", "Tong Zhang", "Jianbing Zhang", "Dongmei Zhang", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87b27d6caaaa3802629fdbf1e7bbe5557ed567f840cee88e45ba7e1fe12c548d", "submitted_at": "2025-06-03T17:59:08Z", "updated_at": "2025-06-03T17:59:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.03143v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.03143", "identifiers": {"arxiv": "2506.03143"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.03143"}], "name": "GUI-Actor: Coordinate-Free Visual Grounding for GUI Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses an attention-based action head and a verifier to select relevant screen regions without generating coordinates as text.", "topics": ["grounding"], "updated_at": "2025-06-03", "url": "https://arxiv.org/abs/2506.03143"}
{"aliases": [], "authors": ["Jiaqi Tang", "Yu Xia", "Yi-Feng Wu", "Yuwei Hu", "Yuhui Chen", "Qing-Guo Chen", "Xiaogang Xu", "Xiangyu Wu", "Hao Lu", "Yanqing Ma", "Shiyin Lu", "Qifeng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b00deba927bda03c079621f70decdeb946ec51a6982531a2086c43033ac9925c", "submitted_at": "2025-06-11T03:43:30Z", "updated_at": "2026-04-21T11:50:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.09373v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.09373", "identifiers": {"arxiv": "2506.09373"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.09373"}], "name": "LPO: Towards Accurate GUI Agent Interaction via Location Preference Optimization", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Location Preference Optimization uses an information-entropy signal to select interaction regions and a distance-based reward to assess click locations. It trains GUI localization through preference optimization and evaluates both offline grounding and online interaction.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-04-21", "url": "https://arxiv.org/abs/2506.09373"}
{"aliases": [], "authors": ["Wenkang Han", "Zhixiong Zeng", "Jing Huang", "Shu Jiang", "Liming Zheng", "Longrong Yang", "Haibo Qiu", "Chang Yao", "Jingyuan Chen", "Lin Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cce7c2e1623ec841e8afed03c2e8e0c6f969f3dd05d53ddc4ef11e74310c21c", "submitted_at": "2025-06-10T12:16:27Z", "updated_at": "2025-11-26T07:48:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.11127v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.11127", "identifiers": {"arxiv": "2506.11127"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.11127"}], "name": "UITron-Speech: Towards Automated GUI Agents Based on Speech Instructions", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-06-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UITron-Speech, which predicts GUI actions from spoken instructions and screenshots. It combines synthetic speech data and mixed-modality training with a training-free grounding refinement step.", "topics": ["grounding", "action-prediction", "data-generation"], "updated_at": "2025-11-26", "url": "https://arxiv.org/abs/2506.11127"}
{"aliases": [], "authors": ["Xingjian Tao", "Yiwei Wang", "Yujun Cai", "Zhicheng Yang", "Jing Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3884b1807cbee971364947aa2f3756937c4c774d9942bb4da3d0fabb69bda07", "submitted_at": "2025-06-18T12:55:35Z", "updated_at": "2025-06-18T12:55:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.15425v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.15425", "identifiers": {"arxiv": "2506.15425"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.15425"}], "name": "Understanding GUI Agent Localization Biases through Logit Sharpness", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a finer-grained taxonomy of GUI localization errors and a logit-based Peak Sharpness Score for analyzing coordinate-prediction uncertainty. A context-aware cropping method refines the visual input without retraining the grounding model.", "topics": ["grounding", "evaluation", "robustness", "training-free"], "updated_at": "2025-06-18", "url": "https://arxiv.org/abs/2506.15425"}
{"aliases": [], "authors": ["Jinjie Wei", "Jiyao Liu", "Lihao Liu", "Ming Hu", "Junzhi Ning", "Mingcheng Li", "Weijie Yin", "Junjun He", "Xiao Liang", "Chao Feng", "Dingkang Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f12d2ae58243b5f89a6a510e8f3ef4073a605c4bafdf63c37fd38db57ea84602", "submitted_at": "2025-06-22T06:30:52Z", "updated_at": "2026-07-21T05:30:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.17913v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.17913", "identifiers": {"arxiv": "2506.17913"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.17913"}], "name": "Learning, Reasoning, Refinement: A Framework for Kahneman's Dual-System Intelligence in GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "CogniGUI combines hierarchical visual parsing of interface elements with a GRPO-based grounding agent that compares alternative interaction paths. Its ScreenSeek benchmark evaluates navigation across applications, changing interface states, and consistency between interfaces.", "topics": ["grounding", "planning", "reinforcement-learning", "evaluation"], "updated_at": "2026-07-21", "url": "https://arxiv.org/abs/2506.17913"}
{"aliases": [], "authors": ["ZongHan Hsieh", "Tzer-Jen Wei", "ShengJing Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb3678b0d0e645a87ce9392b606c1d67dcbc40656880c57d1cbb8ddd8ede7ea1", "submitted_at": "2025-06-30T03:33:02Z", "updated_at": "2025-07-18T06:03:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.23491v3"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.23491", "identifiers": {"arxiv": "2506.23491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.23491"}], "name": "ZonUI-3B: A Lightweight Vision-Language Model for Cross-Resolution GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "ZonUI-3B trains GUI grounding with a curated 24,000-example collection of mobile, desktop, and web screenshots. Its two-stage procedure first learns across platforms and then specializes on high-resolution images, with ablations on data diversity and redundant examples.", "topics": ["grounding", "data-generation", "efficiency"], "updated_at": "2025-07-18", "url": "https://arxiv.org/abs/2506.23491"}
{"aliases": [], "authors": ["Hang Wu", "Hongkai Chen", "Yujun Cai", "Chang Liu", "Qingwen Ye", "Ming-Hsuan Yang", "Yiwei Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e423834b07e74155214597c0fbb66c26b949e2b726568a65617a670a9ab157fa", "submitted_at": "2025-06-12T03:13:21Z", "updated_at": "2025-09-05T17:21:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.00008v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.00008", "identifiers": {"arxiv": "2507.00008"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00008"}], "name": "DiMo-GUI: Advancing Test-time Scaling in GUI Grounding via Modality-Aware Visual Reasoning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates textual and iconic UI elements and iteratively focuses visual regions to refine grounding without additional model training.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-09-05", "url": "https://arxiv.org/abs/2507.00008"}
{"aliases": [], "authors": ["Joonhyung Park", "Peng Tang", "Sagnik Das", "Srikar Appalaraju", "Kunwar Yashraj Singh", "R. Manmatha", "Shabnam Ghadar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aff9b208c446bd57f24258734557e0cd5caaf9146af68e594ec6a7e575919a6f", "submitted_at": "2025-07-08T04:56:57Z", "updated_at": "2025-07-08T04:56:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05673v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.05673", "identifiers": {"arxiv": "2507.05673"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05673"}], "name": "R-VLM: Region-Aware Vision Language Model for Precise GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "R-VLM localizes GUI elements through zoomed-in region proposals and an intersection-over-union-aware training objective. It evaluates the grounding model on ScreenSpot and AgentStudio and studies its effect on navigation in AITW and Mind2Web.", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2025-07-08", "url": "https://arxiv.org/abs/2507.05673"}
{"aliases": [], "authors": ["Yan Yang", "Dongxu Li", "Yutong Dai", "Yuhao Yang", "Ziyang Luo", "Zirui Zhao", "Zhiyuan Hu", "Junzhe Huang", "Amrita Saha", "Zeyuan Chen", "Ran Xu", "Liyuan Pan", "Silvio Savarese", "Caiming Xiong", "Junnan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c1f659a30e1a0bd04219b9383927ce72e5f5b0d20783a89b850036d087155a4", "submitted_at": "2025-07-08T08:52:18Z", "updated_at": "2025-10-03T23:50:19Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2507.05791v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2507.05791", "identifiers": {"arxiv": "2507.05791"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05791"}], "name": "GTA1: GUI Test-time Scaling Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "GTA1 samples multiple action proposals at each GUI step and uses a judge to select among them. A separate grounding model is trained with reinforcement learning to map selected proposals to visual targets, linking planning-time search with click localization.", "topics": ["planning", "grounding", "reinforcement-learning"], "updated_at": "2025-10-03", "url": "https://arxiv.org/abs/2507.05791"}
{"aliases": [], "authors": ["Ziang Ye", "Yang Zhang", "Wentao Shi", "Xiaoyu You", "Fuli Feng", "Tat-Seng Chua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "010a3a54ba37417c97e52434211375447418dd2e7ce1da6c339532534f0a920e", "submitted_at": "2025-07-09T14:36:00Z", "updated_at": "2025-09-24T14:33:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.06899v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.06899", "identifiers": {"arxiv": "2507.06899"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.06899"}], "name": "VisualTrap: A Stealthy Backdoor Attack on GUI Agents via Visual Grounding Manipulation", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-07-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies backdoors that redirect GUI visual grounding even when the task plan is correct. VisualTrap evaluates whether poisoned grounding training transfers to downstream tasks and across web, mobile, and desktop interface settings.", "topics": ["safety-security", "grounding", "evaluation"], "updated_at": "2025-09-24", "url": "https://arxiv.org/abs/2507.06899"}
{"aliases": [], "authors": ["Zihe Yan", "Jiaping Gui", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5db0008a053357b6dba38b55f29642ddbcbf38761c64684fd1725183ff8e56d3", "submitted_at": "2025-07-13T08:36:09Z", "updated_at": "2026-04-07T11:46:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.10610v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.10610", "identifiers": {"arxiv": "2507.10610"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.10610"}], "name": "LaSM: Layer-wise Scaling Mechanism for Defending Pop-up Attack on GUI Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-07-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces LaSM, a training-free defense that selectively scales attention and MLP components to improve alignment with task-relevant screen regions. It evaluates the approach against pop-up-based GUI injection attacks.", "topics": ["safety-security", "grounding", "training-free"], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2507.10610"}
{"aliases": [], "authors": ["Fei Tang", "Zhangxuan Gu", "Zhengxi Lu", "Xuyang Liu", "Shuheng Shen", "Changhua Meng", "Wen Wang", "Wenqi Zhang", "Yongliang Shen", "Weiming Lu", "Jun Xiao", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2a2a49373259cb5903fad8c2fb48a7380732642bedd81a644e735af4f028c63", "submitted_at": "2025-07-21T17:53:42Z", "updated_at": "2025-07-28T16:54:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.15846v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.15846", "identifiers": {"arxiv": "2507.15846"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.15846"}], "name": "GUI-G$^2$: Gaussian Reward Modeling for GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-G2 replaces binary grounding rewards with Gaussian point rewards and spatial-coverage rewards. An adaptive variance tied to element dimensions supplies continuous training feedback for targets of different sizes, evaluated on the ScreenSpot benchmark family.", "topics": ["grounding", "reinforcement-learning", "reward-models"], "updated_at": "2025-07-28", "url": "https://arxiv.org/abs/2507.15846"}
{"aliases": [], "authors": ["Xuehui Wang", "Zhenyu Wu", "JingJing Xie", "Zichen Ding", "Bowen Yang", "Zehao Li", "Zhaoyang Liu", "Qingyun Li", "Xuan Dong", "Zhe Chen", "Weiyun Wang", "Xiangyu Zhao", "Jixuan Chen", "Haodong Duan", "Tianbao Xie", "Chenyu Yang", "Shiqian Su", "Yue Yu", "Yuan Huang", "Yiqian Liu", "Xiao Zhang", "Yanting Zhang", "Xiangyu Yue", "Weijie Su", "Xizhou Zhu", "Wei Shen", "Jifeng Dai", "Wenhai Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b91f7bacbac917e374aa6f656dcfbcb2e220ab3227fa2cff295c63a1c005f9f", "submitted_at": "2025-07-25T17:59:26Z", "updated_at": "2025-07-25T17:59:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.19478v1"}}, "environments": ["web", "desktop", "windows", "macos", "linux", "ios", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.19478", "identifiers": {"arxiv": "2507.19478"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.19478"}], "name": "MMBench-GUI: Hierarchical Multi-Platform Evaluation Framework for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "MMBench-GUI evaluates interface understanding, element grounding, task automation, and collaboration across six named desktop, mobile, and web platforms. It introduces the Efficiency-Quality Area metric for examining execution efficiency alongside task quality in online GUI automation.", "topics": ["evaluation", "grounding", "efficiency", "planning"], "updated_at": "2025-07-25", "url": "https://arxiv.org/abs/2507.19478"}
{"aliases": [], "authors": ["Shuquan Lian", "Yuhang Wu", "Jia Ma", "Yifan Ding", "Zihan Song", "Bingqi Chen", "Xiawu Zheng", "Hui Li", "Rongrong Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "741cce21b25e9b29b99e9804a399695bf8a0ae89df6307d9f9bff68e3e856e85", "submitted_at": "2025-07-29T17:22:07Z", "updated_at": "2026-04-08T16:03:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.22025v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.22025", "identifiers": {"arxiv": "2507.22025"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.22025"}], "name": "UI-AGILE: Advancing GUI Agents with Effective Reinforcement Learning and Precise Inference-Time Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines grounding-oriented reward and sampling improvements with inference-time image decomposition and target selection.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-04-08", "url": "https://arxiv.org/abs/2507.22025"}
{"aliases": [], "authors": ["Miaosen Zhang", "Ziqiang Xu", "Jialiang Zhu", "Qi Dai", "Kai Qiu", "Yifan Yang", "Chong Luo", "Tianyi Chen", "Justin Wagle", "Tim Franklin", "Baining Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a7a32f30d2ada7f739e460429b6453d0e0e56364dda8653b0511ef2e9e6df2a", "submitted_at": "2025-07-31T17:59:09Z", "updated_at": "2025-07-31T17:59:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.23779v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.23779", "identifiers": {"arxiv": "2507.23779"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.23779"}], "name": "Phi-Ground Tech Report: Advancing Perception in GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Reports a GUI-grounding model family and empirical design choices covering source data, perception training, and coordinate prediction.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-07-31", "url": "https://arxiv.org/abs/2507.23779"}
{"aliases": [], "authors": ["Yuhao Cheng", "Liang Tang", "Shuxian Li", "Yukang Huo", "Tiaonan Duan", "Kaer Huang", "Yanzhe Jing", "Yiqiang Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d54ad74695a60cb040d88fcb85549912d7d4ecf6ca2ed78e636d05f8f17d1795", "submitted_at": "2025-08-06T02:57:22Z", "updated_at": "2026-01-22T05:22:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.04037v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.04037", "identifiers": {"arxiv": "2508.04037"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04037"}], "name": "Evolving in Tasks: Empowering the Multi-modality Large Language Model as the Computer Use Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "The Self-Evolution Agent combines automatic generation of verifiable computer-use trajectories with step-wise reinforcement learning for long-horizon training. It also describes a model-enhancement procedure that integrates grounding and planning capabilities without an additional training stage.", "topics": ["data-generation", "reinforcement-learning", "grounding", "planning"], "updated_at": "2026-01-22", "url": "https://arxiv.org/abs/2508.04037"}
{"aliases": [], "authors": ["Weitai Kang", "Bin Lei", "Gaowen Liu", "Caiwen Ding", "Yan Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87b5336c1850c15228880e0625252331b15b79d1cf1e3929abbc01ead70d5f25", "submitted_at": "2025-08-06T12:35:24Z", "updated_at": "2025-08-06T12:35:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04389v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:37Z", "role": "primary", "sha256": "afdfecfc117506609a5bff392254299d88cc84c971469efe0601cbcb02660ecf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2508.04389v1"}], "id": "arxiv:2508.04389", "identifiers": {"arxiv": "2508.04389"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04389"}], "name": "GuirlVG: Incentivize GUI Visual Grounding via Empirical Exploration on Reinforcement Learning", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GuirlVG, an empirical study of reinforcement fine-tuning for GUI visual grounding. It analyzes training components and adds an adaptive KL-based mechanism to limit reward over-optimization.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2508.04389"}
{"aliases": [], "authors": ["Yong Du", "Yuchen Yan", "Fei Tang", "Zhengxi Lu", "Chang Zong", "Weiming Lu", "Shengpei Jiang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7deba3ec19a878cfd402c5dac61e4dca284d478429bc9204264448062299e99", "submitted_at": "2025-08-07T17:54:27Z", "updated_at": "2025-11-13T12:54:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.05615v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.05615", "identifiers": {"arxiv": "2508.05615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.05615"}], "name": "Test-Time Reinforcement Learning for GUI Grounding via Region Consistency", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-RC aggregates multiple grounding predictions into spatial voting grids to choose a consensus region. GUI-RCPO then turns agreement with that region into a reward for test-time policy updates on unlabeled GUI screenshots.", "topics": ["grounding", "reinforcement-learning", "training-free"], "updated_at": "2025-11-13", "url": "https://arxiv.org/abs/2508.05615"}
{"aliases": [], "authors": ["Yuhang Liu", "Zeyu Liu", "Shuanghe Zhu", "Pengxiang Li", "Congkai Xie", "Jiasheng Wang", "Xavier Hu", "Xiaotian Han", "Jianbo Yuan", "Xinyao Wang", "Shengyu Zhang", "Hongxia Yang", "Fei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a69d2bdb4d01e6e5c9f08da33281c8247f59af52f3bc08dfed6b7e24da88e194", "submitted_at": "2025-08-07T17:49:56Z", "updated_at": "2025-12-08T10:49:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.05731v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.05731", "identifiers": {"arxiv": "2508.05731"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.05731"}], "name": "InfiGUI-G1: Advancing GUI Grounding with Adaptive Exploration Policy Optimization", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses multiple candidate answers and adaptive exploration rewards to improve spatial and semantic alignment in GUI grounding.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-12-08", "url": "https://arxiv.org/abs/2508.05731"}
{"aliases": [], "authors": ["Fengxian Ji", "Jingpu Yang", "Zirui Song", "Yuanxi Wang", "Zhexuan Cui", "Yuke Li", "Qian Jiang", "Miao Fang", "Xiuying Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8c91a1f85602760c706e56426c0d2d870bbc101ca500aa3c7c391938c3f1676", "submitted_at": "2025-08-12T15:12:42Z", "updated_at": "2025-08-12T15:12:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.09241v1"}}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.09241", "identifiers": {"arxiv": "2508.09241"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.09241"}], "name": "FineState-Bench: A Comprehensive Benchmark for Fine-Grained State Control in GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "FineState-Bench evaluates fine-grained GUI perception and control across desktop, web, and mobile interfaces. Its Visual Diagnostic Assistant separates perception and localization contributions in controlled interaction experiments.", "topics": ["evaluation", "grounding", "observability", "screen-understanding"], "updated_at": "2025-08-12", "url": "https://arxiv.org/abs/2508.09241"}
{"aliases": [], "authors": ["Zhangxuan Gu", "Zhengwen Zeng", "Zhenyu Xu", "Xingran Zhou", "Shuheng Shen", "Yunfei Liu", "Beitong Zhou", "Changhua Meng", "Tianyu Xia", "Weizhi Chen", "Yue Wen", "Jingya Dou", "Fei Tang", "Jinzhen Lin", "Yulin Liu", "Zhenlin Guo", "Yichen Gong", "Heng Jia", "Changlong Gao", "Yuan Guo", "Yong Deng", "Zhenyu Guo", "Liang Chen", "Weiqiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db88c4ba15c48210f0b618a8b5d2b9c02184f2e4375d45adf22b9a0a395abc45", "submitted_at": "2025-08-14T16:58:07Z", "updated_at": "2025-08-15T14:49:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.10833v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2508.10833", "identifiers": {"arxiv": "2508.10833"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.10833"}], "name": "UI-Venus Technical Report: Building High-performance UI Agents with RFT", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-08-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies screenshot-based GUI grounding and navigation using curated interaction data, reward design, and reinforcement fine-tuning.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-08-15", "url": "https://arxiv.org/abs/2508.10833"}
{"aliases": [], "authors": ["Jikai Chen", "Long Chen", "Dong Wang", "Qinglin Su", "Zhixuan Chu", "Bingguang Hao", "Leilei Gan", "Chenyi Zhuang", "Jinjie Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfb50e10184cded6c18d4c49401145bc7cfa00b0ba21bce964abe8f459af5626", "submitted_at": "2025-08-19T08:47:44Z", "updated_at": "2026-01-20T06:39:19Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.13634v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.13634", "identifiers": {"arxiv": "2508.13634"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.13634"}], "name": "V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "V2P trains GUI grounding by suppressing attention to background regions and emphasizing the center of the target element. It represents target importance with size-dependent Gaussian heatmaps and evaluates the contribution of each component on ScreenSpot-v2 and ScreenSpot-Pro.", "topics": ["grounding", "screen-understanding"], "updated_at": "2026-01-20", "url": "https://arxiv.org/abs/2508.13634"}
{"aliases": [], "authors": ["Tanvir Bhathal", "Asanshay Gupta"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67da98d98fec647de9b482fabd8c7861063c0fdb91f1ec303fe8df759ee0f69d", "submitted_at": "2025-08-23T11:02:59Z", "updated_at": "2025-08-23T11:02:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.16987v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.16987", "identifiers": {"arxiv": "2508.16987"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.16987"}], "name": "WebSight: A Vision-First Architecture for Robust Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebSight uses a vision-language model trained on web-interface examples within a modular agent with planning, action, verification, and episodic memory components. It evaluates element clicking on Showdown Clicks and full browsing tasks on WebVoyager using visual observations.", "topics": ["web-navigation", "grounding", "planning", "memory"], "updated_at": "2025-08-23", "url": "https://arxiv.org/abs/2508.16987"}
{"aliases": [], "authors": ["Zhixiong Zeng", "Jing Huang", "Liming Zheng", "Wenkang Han", "Yufeng Zhong", "Lei Chen", "Longrong Yang", "Yingjie Chu", "Yuzhi He", "Lin Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b2a14819c7d7e8258947c9db8cc89012f4d133eaced4e4b904c3bfd1761181e5", "submitted_at": "2025-08-29T16:40:57Z", "updated_at": "2025-08-29T16:40:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.21767v1"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.21767", "identifiers": {"arxiv": "2508.21767"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.21767"}], "name": "UItron: Foundational GUI Agent with Advanced Perception and Planning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "UItron combines GUI perception and planning data engineering with supervised training and curriculum reinforcement learning in connected mobile and PC environments. It also collects more than one million operation steps across 100 Chinese apps and builds offline and online evaluation environments.", "topics": ["data-generation", "reinforcement-learning", "planning", "grounding"], "updated_at": "2025-08-29", "url": "https://arxiv.org/abs/2508.21767"}
{"aliases": [], "authors": ["Jingru Fan", "Yufan Dang", "Jingyao Wu", "Huatao Li", "Runde Yang", "Xiyuan Yang", "Yuheng Wang", "Chen Qian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49a887b12b3395821aa100a66e61aac808309ef947ae752475df3d2db9a4fcd8", "submitted_at": "2025-09-02T15:48:21Z", "updated_at": "2025-10-17T00:57:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.02444v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:33.619660+00:00", "role": "primary", "sha256": "6c2591dad8893f87cffdea211584a3733dc2137dafa9416c1885d1b0126fd540", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2509.02444v2"}], "id": "arxiv:2509.02444", "identifiers": {"arxiv": "2509.02444"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:10.946679+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.02444"}], "name": "AppCopilot: Toward General, Accurate, Long-Horizon, and Efficient Mobile Agent", "notes": "Trace reuse assumes stable layouts and fixed tasks. The described function-call routes are user-selected, and coordinate calibration depends on valid detected widgets. No particular operating-system coverage, latency guarantee or reproduced result is asserted. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (4.2 OCR/OR Region Localization; 4.2.2 Coordinate Calibration; 4.3 Multi-Agent Voting; 5.2 Task Decomposition; 6.2 Experience Reuse; 6.3 Function Calling); not an entire-paper reading or reproduction claim.", "published_at": "2025-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a mobile-agent stack combining region-based coordinate correction, hierarchical task planning, multi-agent voting, and reuse of successful interaction traces.", "topics": ["planning", "grounding", "workflow", "mobile-navigation", "long-horizon"], "updated_at": "2025-10-17", "url": "https://arxiv.org/abs/2509.02444"}
{"aliases": [], "authors": ["Wanfu Wang", "Qipeng Huang", "Guangquan Xue", "Xiaobo Liang", "Juntao Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "875e2452985bae9e67d7071121ab13f5205307ecc30bf33f237048a44e75b3bc", "submitted_at": "2025-09-04T14:17:01Z", "updated_at": "2025-09-04T14:17:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04243v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.04243", "identifiers": {"arxiv": "2509.04243"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.04243"}], "name": "Learning Active Perception via Self-Evolving Preference Optimization for GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "LASER builds preference data for multi-step GUI perception by combining Monte Carlo quality estimates with region-overlap evaluation. Its self-evolving training encourages relevant region selection and adapts the amount of reasoning before predicting screen coordinates.", "topics": ["grounding", "data-generation", "screen-understanding"], "updated_at": "2025-09-04", "url": "https://arxiv.org/abs/2509.04243"}
{"aliases": [], "authors": ["Hongyi Jing", "Jiafu Chen", "Chen Rao", "Ziqiang Dang", "Jiajie Teng", "Tianyi Chu", "Juncheng Mo", "Shuo Fang", "Huaizhong Lin", "Rui Lv", "Chenguang Ma", "Lei Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f74081d31c98d18e10a5febc5de436a3788fc4156215e4be8fc54b1c53260f58", "submitted_at": "2025-09-05T08:24:12Z", "updated_at": "2025-09-05T08:24:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04908v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:46Z", "role": "primary", "sha256": "e4304967a9a63ead71f00e1933965321540df913f534c0711cc18b9190560cde", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.04908v1"}], "id": "arxiv:2509.04908", "identifiers": {"arxiv": "2509.04908"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.04908"}], "name": "SparkUI-Parser: Enhancing GUI Perception with Robust Grounding and Parsing", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-09-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SparkUI-Parser, which adds continuous coordinate decoding and a rejection mechanism for nonexistent targets to a multimodal model. It also presents ScreenParse for evaluating structural interface perception.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2025-09-05", "url": "https://arxiv.org/abs/2509.04908"}
{"aliases": [], "authors": ["Weiming Li", "Yan Shao", "Jing Yang", "Yujing Lu", "Ling Zhong", "Yuhan Wang", "Min Yu", "Tongxiao Ruan", "Manni Duan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75ffc21abebbc2c362798f3455b9e6240b58407572e5633d5171e5612cfae71b", "submitted_at": "2025-09-15T03:28:29Z", "updated_at": "2026-06-10T01:13:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.11548v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.11548", "identifiers": {"arxiv": "2509.11548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11548"}], "name": "How Auxiliary Reasoning Unleashes GUI Grounding in VLMs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies the gap between latent visual grounding and explicit coordinate output in general vision-language models. Three zero-shot methods add axes, grids, or labeled intersections to GUI screenshots so models can express spatial locations without fine-tuning.", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-06-10", "url": "https://arxiv.org/abs/2509.11548"}
{"aliases": [], "authors": ["Zongru Wu", "Rui Mao", "Zhiyuan Tian", "Pengzhou Cheng", "Tianjie Ju", "Zheng Wu", "Lingzhong Dong", "Haiyue Sheng", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "be8620b021bd9218c1143282f1c7a7ea06be2abe6fc0928a310cfc8ac3172a18", "submitted_at": "2025-09-17T01:14:14Z", "updated_at": "2026-03-18T08:12:54Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2509.13615v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:53Z", "role": "primary", "sha256": "28426488ebca15b74a693ea9fcb2b0ce3165d3baf97b76a4884b34a0d1997af4", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.13615v5"}], "id": "arxiv:2509.13615", "identifiers": {"arxiv": "2509.13615"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:53Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.13615"}], "name": "See, Think, Act: Teaching Multimodal Agents to Effectively Interact with GUI by Identifying Toggles", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a benchmark for binary GUI-toggle instructions and State-aware Reasoning, which compares the observed control state with the requested state before choosing an action. It evaluates agents on toggle tasks and broader interactive benchmarks.", "topics": ["state-tracking", "grounding", "evaluation"], "updated_at": "2026-03-18", "url": "https://arxiv.org/abs/2509.13615"}
{"aliases": [], "authors": ["Xianhang Ye", "Yiqing Li", "Wei Dai", "Miancan Liu", "Ziyuan Chen", "Zhangye Han", "Hongbo Min", "Jinkui Ren", "Xiantao Zhang", "Wen Yang", "Zhi Jin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e415074c6a4b4995350d6936d338682b9a4469f1cbed3a83e0ac44eee6f02bef", "submitted_at": "2025-09-19T02:34:50Z", "updated_at": "2025-09-19T02:34:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.15532v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.15532", "identifiers": {"arxiv": "2509.15532"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15532"}], "name": "GUI-ARP: Enhancing Grounding with Adaptive Region Perception for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-ARP uses visual attention to crop relevant screenshot regions and decides whether a target needs single-stage or multi-stage analysis. Supervised fine-tuning followed by GRPO trains this adaptive procedure for high-resolution GUI grounding.", "topics": ["grounding", "reinforcement-learning", "screen-understanding"], "updated_at": "2025-09-19", "url": "https://arxiv.org/abs/2509.15532"}
{"aliases": [], "authors": ["Shaojie Zhang", "Ruoceng Zhang", "Pei Fu", "Shaokang Wang", "Jiahui Yang", "Xin Du", "Shiqi Cui", "Bin Qin", "Ying Huang", "Zhenbo Luo", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d190eefae9309d87f4ea4780d3483ae0ebd834e0bf70e7f2af204c3788d87e7e", "submitted_at": "2025-09-19T04:03:44Z", "updated_at": "2025-10-27T11:34:58Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2509.15566v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.15566", "identifiers": {"arxiv": "2509.15566"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15566"}], "name": "BTL-UI: Blink-Think-Link Reasoning Model for GUI Agent", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "BTL-UI separates GUI interaction into screen-region attention, planning and executable-command generation. It adds an automated annotation pipeline and rule-based process/outcome rewards for training, with evaluation on both static understanding and dynamic interaction.", "topics": ["grounding", "planning", "data-generation", "reinforcement-learning"], "updated_at": "2025-10-27", "url": "https://arxiv.org/abs/2509.15566"}
{"aliases": [], "authors": ["Hongxin Li", "Jingran Su", "Jingfan Chen", "Zheng Ju", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97b7c0690ba7deb7dbbdac6ac94dd82e3864c440cb4325bf143ad735919bb08b", "submitted_at": "2025-09-22T03:04:53Z", "updated_at": "2025-09-22T03:04:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.17328v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.17328", "identifiers": {"arxiv": "2509.17328"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.17328"}], "name": "UIPro: Unleashing Superior Interaction Capability For GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "UIPro pretrains GUI understanding on 20.6 million tasks and then harmonizes interaction datasets through a unified action space. Continued fine-tuning uses the merged multi-platform data to train action prediction for generalist GUI operation.", "topics": ["data-generation", "action-prediction", "grounding", "imitation-learning"], "updated_at": "2025-09-22", "url": "https://arxiv.org/abs/2509.17328"}
{"aliases": [], "authors": ["Yu Zhao", "Wei-Ning Chen", "Huseyin Atahan Inan", "Samuel Kessler", "Lu Wang", "Lukas Wutschitz", "Fangkai Yang", "Chaoyun Zhang", "Pasquale Minervini", "Saravan Rajmohan", "Robert Sim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83fa51d3642fa7786f7f6c44fcf5d5ab5876d7ae26e5f6184ddff4d5e2009181", "submitted_at": "2025-09-25T20:38:01Z", "updated_at": "2026-05-25T20:58:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21552v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.21552", "identifiers": {"arxiv": "2509.21552"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21552"}], "name": "Learning GUI Grounding with Spatial Reasoning from Visual Feedback", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Cursor treats GUI grounding as a sequence of cursor movements guided by visible feedback and movement history. Online reinforcement learning with trajectory rewards trains the model to refine the target location across multiple steps.", "topics": ["grounding", "reinforcement-learning", "action-prediction", "robustness"], "updated_at": "2026-05-25", "url": "https://arxiv.org/abs/2509.21552"}
{"aliases": [], "authors": ["Zhecheng Li", "Guoxian Song", "Yiwei Wang", "Zhen Xiong", "Junsong Yuan", "Yujun Cai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b42e784de8dd007c60d234b7331bd98b5a04c58c13696bf436c81c7eccee37f", "submitted_at": "2025-09-29T00:06:31Z", "updated_at": "2025-09-29T00:06:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.24133v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.24133", "identifiers": {"arxiv": "2509.24133"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.24133"}], "name": "Generalist Scanner Meets Specialist Locator: A Synergistic Coarse-to-Fine Framework for Robust GUI Grounding", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "GMS pairs a general vision-language scanner that proposes interface regions with a specialist locator that predicts coordinates within them. Hierarchical coarse-to-fine search coordinates the two components for GUI grounding on ScreenSpot-Pro.", "topics": ["grounding", "planning", "evaluation"], "updated_at": "2025-09-29", "url": "https://arxiv.org/abs/2509.24133"}
{"aliases": [], "authors": ["Zhen Yang", "Zi-Yi Dou", "Di Feng", "Forrest Huang", "Anh Nguyen", "Keen You", "Omar Attia", "Yuhao Yang", "Michael Feng", "Haotian Zhang", "Ram Ramrakhya", "Chao Jia", "Jeffrey Nichols", "Alexander Toshev", "Yinfei Yang", "Zhe Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "daa35e6bcd86fbdb709652d3592776ed93aebe2a0ebb3595fed91478dc0e3e0a", "submitted_at": "2025-09-30T17:13:56Z", "updated_at": "2025-09-30T17:13:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.26539v1"}}, "environments": ["android", "web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.26539", "identifiers": {"arxiv": "2509.26539"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.26539"}], "name": "Ferret-UI Lite: Lessons from Building Small On-Device GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Ferret-UI Lite develops a 3B GUI agent using mixed real and synthetic data, reasoning with visual tools at inference, and reinforcement learning. It evaluates grounding and navigation across mobile, web, and desktop settings, including AndroidWorld and OSWorld.", "topics": ["grounding", "reinforcement-learning", "tool-integration", "efficiency"], "updated_at": "2025-09-30", "url": "https://arxiv.org/abs/2509.26539"}
{"aliases": [], "authors": ["Suyuchen Wang", "Tianyu Zhang", "Ahmed Masry", "Christopher Pal", "Spandana Gella", "Bang Liu", "Perouz Taslakian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f0c7eedadb9a321b222d4081271d6c979a25e1f64714cea8cf2ff088c5e61df", "submitted_at": "2025-10-03T17:59:34Z", "updated_at": "2025-10-03T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.03230v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.03230", "identifiers": {"arxiv": "2510.03230"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.03230"}], "name": "Improving GUI Grounding with Explicit Position-to-Coordinate Mapping", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "This work addresses GUI localization errors caused by mismatches between visual patch positions and output pixel coordinates. RULER tokens provide explicit coordinate references, while Interleaved MRoPE balances width and height encoding; evaluation uses the ScreenSpot benchmark family.", "topics": ["grounding", "screen-understanding", "robustness"], "updated_at": "2025-10-03", "url": "https://arxiv.org/abs/2510.03230"}
{"aliases": [], "authors": ["Bin Lei", "Nuo Xu", "Ali Payani", "Mingyi Hong", "Chunhua Liao", "Yu Cao", "Caiwen Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "798fc5ca4fe483a1235f0979b0c641e9c1732660780c13340c5221319a1b28fc", "submitted_at": "2025-10-05T05:15:45Z", "updated_at": "2025-10-05T05:15:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.04039v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.04039", "identifiers": {"arxiv": "2510.04039"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.04039"}], "name": "\\textsc{GUI-Spotlight}: Adaptive Iterative Focus Refinement for Enhanced GUI Visual Grounding", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Spotlight trains a grounding model to reason over images and call tools that iteratively narrow its visual focus. The paper evaluates this refinement method on ScreenSpot-Pro after training on 18,500 samples.", "topics": ["grounding", "planning", "tool-integration"], "updated_at": "2025-10-05", "url": "https://arxiv.org/abs/2510.04039"}
{"aliases": [], "authors": ["Ho Fai Leung", "Xiaoyan Xi", "Fei Zuo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33b21a63d51148f2b5c30478c302e3df9b92e03e2e106a1c68224f2c835a8608", "submitted_at": "2025-10-21T10:11:33Z", "updated_at": "2025-10-21T10:11:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.18488v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.18488", "identifiers": {"arxiv": "2510.18488"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.18488"}], "name": "AndroidControl-Curated: Revealing the True Potential of GUI Agents through Benchmark Purification", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "AndroidControl-Curated audits ambiguities and factual errors in AndroidControl and produces a refined evaluation dataset. The paper also studies post-training a small Magma-R1 model on a curated subset to examine the effect of data quality on GUI action assessment.", "topics": ["evaluation", "data-generation", "grounding"], "updated_at": "2025-10-21", "url": "https://arxiv.org/abs/2510.18488"}
{"aliases": [], "authors": ["Liangyu Chen", "Hanzhang Zhou", "Chenglin Cai", "Jianan Zhang", "Panrong Tong", "Quyu Kong", "Xu Zhang", "Chen Liu", "Yuqi Liu", "Wenxuan Wang", "Yue Wang", "Qin Jin", "Steven Hoi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3378dd9187f8c8df983d97b10b2ca5b64b2e2332760a734a09db8ce34a398a0b", "submitted_at": "2025-10-23T07:18:32Z", "updated_at": "2025-10-23T07:18:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.20286v1"}}, "environments": ["cross-platform", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.20286", "identifiers": {"arxiv": "2510.20286"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.20286"}], "name": "UI-Ins: Enhancing GUI Grounding with Multi-Perspective Instruction-as-Reasoning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Treats alternative instruction formulations as reasoning paths and trains GUI-grounding models to select and combine useful perspectives.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-10-23", "url": "https://arxiv.org/abs/2510.20286"}
{"aliases": [], "authors": ["Shaojie Zhang", "Pei Fu", "Ruoceng Zhang", "Jiahui Yang", "Anan Du", "Xiuwen Xi", "Shaokang Wang", "Ying Huang", "Bin Qin", "Zhenbo Luo", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6869e86d7cd4242ba8dcf942bb1cef3c157dff4f0ef37f7138ff5025d082c7b8", "submitted_at": "2025-10-31T08:07:02Z", "updated_at": "2026-05-27T02:32:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.27266v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.27266", "identifiers": {"arxiv": "2510.27266"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.27266"}], "name": "Enhancing Trustworthy GUI Grounding via Self-Critiqued Reinforcement Learning", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "HyperClick trains GUI grounders to produce a click and an accompanying confidence estimate using rewards for correctness and confidence alignment. Its evaluation examines localization, calibration, and abstention behavior.", "topics": ["grounding", "reinforcement-learning", "robustness", "evaluation"], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2510.27266"}
{"aliases": [], "authors": ["Shijie Zhou", "Viet Dac Lai", "Hao Tan", "Jihyung Kil", "Wanrong Zhu", "Changyou Chen", "Ruiyi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c2224c58c8fbac7a68f3757391f5e6a4f62c32fa42aae86bf9821c0755b402b1", "submitted_at": "2025-11-02T05:34:21Z", "updated_at": "2026-06-30T10:43:39Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2511.00810v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.00810", "identifiers": {"arxiv": "2511.00810"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.00810"}], "name": "GUI-AIMA: Aligning Intrinsic Multimodal Attention with a Context Anchor for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-AIMA, which supervises instruction-relevant attention over visual patches instead of directly training coordinate text generation. Patch-level grounding can be combined with an optional zoom stage and is evaluated across GUI localization benchmarks.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2511.00810"}
{"aliases": [], "authors": ["Jian Mu", "Chaoyun Zhang", "Chiming Ni", "Lu Wang", "Bo Qiao", "Kartik Mathur", "Qianhui Wu", "Yuhang Xie", "Xiaojun Ma", "Mengyu Zhou", "Si Qin", "Liqun Li", "Yu Kang", "Minghua Ma", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b063a4d2b3a552aa05a153745978473a7d165f2c6b4b97e4c91eff11b7b0f80", "submitted_at": "2025-11-06T12:19:02Z", "updated_at": "2025-11-10T12:27:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.04307v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.04307", "identifiers": {"arxiv": "2511.04307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.04307"}], "name": "GUI-360$^\\circ$: A Comprehensive Dataset and Benchmark for Computer-Using Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-11-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Collects Windows application trajectories with screenshots and action traces for grounding, screen parsing, and action prediction.", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-11-10", "url": "https://arxiv.org/abs/2511.04307"}
{"aliases": [], "authors": ["Xiangwu Guo", "Difei Gao", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a5f43873a47e5b47250cedd3d860abe2dc5ec9b85fe657444bf2b7378f3e8f7", "submitted_at": "2025-11-09T15:13:45Z", "updated_at": "2025-11-09T15:13:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.06417v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.06417", "identifiers": {"arxiv": "2511.06417"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.06417"}], "name": "AUTO-Explorer: Automated Data Collection for GUI Agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Auto-Explorer, an automated GUI exploration method that records software states for training-data collection. UIXplore evaluates exploration coverage, and collected data is used to fine-tune and test GUI element grounding.", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-11-09", "url": "https://arxiv.org/abs/2511.06417"}
{"aliases": [], "authors": ["Aarash Feizi", "Shravan Nayak", "Xiangru Jian", "Kevin Qinghong Lin", "Kaixin Li", "Rabiul Awal", "Xing Han Lù", "Johan Obando-Ceron", "Juan A. Rodriguez", "Nicolas Chapados", "David Vazquez", "Adriana Romero-Soriano", "Reihaneh Rabbany", "Perouz Taslakian", "Christopher Pal", "Spandana Gella", "Sai Rajeswar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f2b80f294ddf5e3331dcb47a836864032f88f2fdbe35588bfef000cd9ceddc8", "submitted_at": "2025-11-10T17:35:21Z", "updated_at": "2026-06-09T21:30:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.07332v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.07332", "identifiers": {"arxiv": "2511.07332"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.07332"}], "name": "Grounding Computer Use Agents on Human Demonstrations", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "GroundCUA turns expert desktop demonstrations into densely annotated screenshots and grounding instructions. The GroundNext model family uses this supervision and subsequent reinforcement learning, with both grounding benchmarks and an OSWorld agent configuration used for evaluation.", "topics": ["grounding", "data-generation", "evaluation", "reinforcement-learning"], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2511.07332"}
{"aliases": [], "authors": ["Yuan Zhao", "Hualei Zhu", "Tingyu Jiang", "Shen Li", "Xiaohang Xu", "Hao Henry Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "101d38ed439c41ebfa826e990697b502b9126728956295d25b0ea9d539ff23e3", "submitted_at": "2025-11-13T03:41:02Z", "updated_at": "2025-11-13T03:41:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.10705v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.10705", "identifiers": {"arxiv": "2511.10705"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.10705"}], "name": "Co-EPG: A Framework for Co-Evolution of Planning and Grounding in Autonomous GUI Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Co-EPG alternates reinforcement learning for a planner with updates to a grounding model using generated training examples. The evolving grounder provides feedback to the planner, and evaluation uses Multimodal Mind2Web and AndroidControl.", "topics": ["grounding", "planning", "reinforcement-learning", "data-generation"], "updated_at": "2025-11-13", "url": "https://arxiv.org/abs/2511.10705"}
{"aliases": [], "authors": ["SeokJoo Kwak", "Jihoon Kim", "Boyoun Kim", "Jung Jae Yoon", "Wooseok Jang", "Jeonghoon Hong", "Jaeho Yang", "Yeong-Dae Kwon"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cdabdd4fc56b95c2b0c801fefd76d4c724aa0a1d7f762623d4314ca9a38e85c", "submitted_at": "2025-11-17T07:38:05Z", "updated_at": "2025-11-17T07:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.13087v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2511.13087", "identifiers": {"arxiv": "2511.13087"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.13087"}], "name": "MEGA-GUI: Multi-stage Enhanced Grounding Agents for GUI Elements", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MEGA-GUI, which separates coarse region selection from fine-grained interface-element grounding. Bidirectional zooming and context-aware instruction rewriting coordinate specialized vision-language agents across visual scales.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-11-17", "url": "https://arxiv.org/abs/2511.13087"}
{"aliases": [], "authors": ["Jiayuan Zhang", "Kaiquan Chen", "Zhihao Lu", "Enshen Zhou", "Qian Yu", "Jing Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af9da8a618674b5a40f95029c4797b48a850e0a99de827d3e526b21797f8eef4", "submitted_at": "2025-11-26T13:49:39Z", "updated_at": "2025-11-26T13:49:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.21398v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.21398", "identifiers": {"arxiv": "2511.21398"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.21398"}], "name": "Prune4Web: DOM Tree Pruning Programming for Web Agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Prune4Web, where generated scoring programs filter DOM elements according to decomposed subtasks before grounding. An annotation pipeline and two-turn training procedure coordinate the planner, programmatic filter and grounder.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2025-11-26", "url": "https://arxiv.org/abs/2511.21398"}
{"aliases": [], "authors": ["Neeraj Anand", "Rishabh Jain", "Sohan Patnaik", "Balaji Krishnamurthy", "Mausoom Sarkar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "faad266b89aed9b210a813b69184d7236f9098c0e095cc426402e154cbdd067b", "submitted_at": "2025-11-30T11:32:54Z", "updated_at": "2025-12-11T18:16:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.00846v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.00846", "identifiers": {"arxiv": "2512.00846"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.00846"}], "name": "AFRAgent : An Adaptive Feature Renormalization Based High Resolution Aware GUI agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AFRAgent, which uses adaptive feature renormalization to combine low-resolution visual embeddings with high-resolution details. The architecture is evaluated for smartphone GUI automation on Meta-GUI and AITW.", "topics": ["grounding", "mobile-navigation", "efficiency"], "updated_at": "2025-12-11", "url": "https://arxiv.org/abs/2512.00846"}
{"aliases": [], "authors": ["Hyunjun Kim", "Sooyoung Ryu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e445d0ca79f727a6c1d51e83381c3fd1d8e1fddeff7de99fe70ca06bc1bff49", "submitted_at": "2025-12-01T01:18:21Z", "updated_at": "2025-12-01T01:18:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.01174v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2512.01174", "identifiers": {"arxiv": "2512.01174"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.01174"}], "name": "DrawingBench: Evaluating Spatial Reasoning and UI Interaction Capabilities of Large Language Models through Mouse-Based Drawing Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "DrawingBench evaluates sequences of low-level mouse-based GUI actions through 250 drawing prompts and rule-based criteria. Action inspection and multi-turn external feedback support analysis of spatial reasoning, tool-state management, and refinement.", "topics": ["evaluation", "grounding", "human-in-the-loop", "state-tracking"], "updated_at": "2025-12-01", "url": "https://arxiv.org/abs/2512.01174"}
{"aliases": [], "authors": ["Aiden Yiliu Li", "Bizhi Yu", "Daoan Lei", "Tianhe Ren", "Shilong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4868c3c85545532e9c208623596d3810c85e32e3cf11429a68847ac11c2ba1cb", "submitted_at": "2025-12-01T18:37:19Z", "updated_at": "2025-12-01T18:37:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.01979v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.01979", "identifiers": {"arxiv": "2512.01979"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.01979"}], "name": "Chain-of-Ground: Improving GUI Grounding via Iterative Reasoning and Reference Feedback", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Chain-of-Ground uses iterative visual reasoning and reference feedback to refine GUI localization without additional training. It evaluates on ScreenSpot-Pro and introduces TPanel-UI to probe grounding on industrial control panels with visual distortions.", "topics": ["grounding", "training-free", "robustness", "evaluation"], "updated_at": "2025-12-01", "url": "https://arxiv.org/abs/2512.01979"}
{"aliases": [], "authors": ["Zhiyuan Jiang", "Shenghao Xie", "Wenyi Li", "Wenqiang Zu", "Peihang Li", "Jiahao Qiu", "Siqi Pei", "Lei Ma", "Tiejun Huang", "Mengdi Wang", "Shilong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bac971bd0e7cd946f81e28d5976eb933cbb95b97ca846ee624a0c36b48760de7", "submitted_at": "2025-12-05T18:39:12Z", "updated_at": "2025-12-05T18:39:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.05941v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.05941", "identifiers": {"arxiv": "2512.05941"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.05941"}], "name": "Zoom in, Click out: Unlocking and Evaluating the Potential of Zooming for GUI Grounding", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "ZoomClick uses repeated zooming and context adjustment to improve GUI grounding without training. The paper studies zoom depth, crop sizes, and related choices and introduces GUIZoom-Bench to evaluate how grounding models respond to zoomed views.", "topics": ["grounding", "training-free", "evaluation", "screen-understanding"], "updated_at": "2025-12-05", "url": "https://arxiv.org/abs/2512.05941"}
{"aliases": [], "authors": ["Yunzhu Zhang", "Zeyu Pan", "Zhengwen Zeng", "Shuheng Shen", "Changhua Meng", "Linchao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95f6e22cda1084d32bbd3f102ea780bfbf392ee578b7c20542314ef76618a56b", "submitted_at": "2025-12-09T12:19:00Z", "updated_at": "2025-12-09T12:19:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.08529v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.08529", "identifiers": {"arxiv": "2512.08529"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.08529"}], "name": "MVP: Multiple View Prediction Improves GUI Grounding", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "MVP addresses unstable GUI coordinates by predicting across several attention-guided cropped views. It clusters the resulting coordinates and selects the densest cluster centroid, evaluating this training-free inference method across grounding models.", "topics": ["grounding", "training-free", "robustness"], "updated_at": "2025-12-09", "url": "https://arxiv.org/abs/2512.08529"}
{"aliases": [], "authors": ["Beitong Zhou", "Zhexiao Huang", "Yuan Guo", "Zhangxuan Gu", "Tianyu Xia", "Zichen Luo", "Fei Tang", "Dehan Kong", "Yanyi Shang", "Suling Ou", "Zhenlin Guo", "Changhua Meng", "Shuheng Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bc8a1da3993ead9b36af47a9332fe9a1a2427101ad0d91da654b4a4b25141df", "submitted_at": "2025-12-18T13:09:09Z", "updated_at": "2025-12-18T13:09:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.16501v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2512.16501", "identifiers": {"arxiv": "2512.16501"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.16501"}], "name": "VenusBench-GD: A Comprehensive Multi-Platform GUI Benchmark for Diverse Grounding Tasks", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "VenusBench-GD is a bilingual, cross-platform GUI grounding benchmark organized into basic and advanced tasks. Its annotation pipeline and six-subtask taxonomy support comparisons of general multimodal models and GUI-specialized models across different grounding demands.", "topics": ["grounding", "evaluation", "data-generation", "robustness"], "updated_at": "2025-12-18", "url": "https://arxiv.org/abs/2512.16501"}
{"aliases": [], "authors": ["Brandon Huang", "Hang Hua", "Zhuoran Yu", "Trevor Darrell", "Rogerio Feris", "Roei Herzig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37cd0aa99ed8bad167e25f6e2b48e7cd1cf01249508555f657215213ee94a301", "submitted_at": "2025-12-19T04:09:24Z", "updated_at": "2025-12-31T17:30:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.17221v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.17221", "identifiers": {"arxiv": "2512.17221"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.17221"}], "name": "DAVE: A VLM Vision Encoder for Document Understanding and Web Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "DAVE trains a vision encoder for document images and web agents using self-supervised pretraining followed by supervised parsing and localization. It combines decoder-specific encoder merging with feature fusion and evaluates web localization and agent benchmarks alongside document tasks.", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2025-12-31", "url": "https://arxiv.org/abs/2512.17221"}
{"aliases": [], "authors": ["Sarthak Mehrotra", "Sairam V C Rebbapragada", "Mani Hemanth Reddy Bonthu", "Vineeth N Balasubramanian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "041466768673c1ea49bceda2ad08784c0fa0e896bcee5452a1e41d51ac3b81d8", "submitted_at": "2025-12-26T12:09:15Z", "updated_at": "2025-12-26T12:09:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.22009v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.22009", "identifiers": {"arxiv": "2512.22009"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.22009"}], "name": "iSHIFT: Lightweight Slow-Fast GUI Agent with Adaptive Perception", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-12-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces iSHIFT, which combines latent reasoning with perception-control tokens to adapt GUI inference. The agent switches between detailed regional grounding and faster use of global visual cues.", "topics": ["efficiency", "grounding", "screen-understanding"], "updated_at": "2025-12-26", "url": "https://arxiv.org/abs/2512.22009"}
{"aliases": ["ShowUI-pi", "ShowUI Pi"], "authors": ["Siyuan Hu", "Kevin Qinghong Lin", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b7d1f358b450f2a6c3deb00f1b16c54d443f48fda3a7dc70c77456bebf35ad7c", "submitted_at": "2025-12-31T16:51:14Z", "updated_at": "2025-12-31T16:51:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.24965v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2512.24965", "identifiers": {"arxiv": "2512.24965"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.24965"}], "name": "ShowUI-$π$: Flow-based Generative Models as GUI Dexterous Hands", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Models discrete clicks and continuous drags with flow-based action generation, and introduces drag trajectories and the ScreenDrag benchmark.", "topics": ["grounding", "evaluation", "data-generation"], "updated_at": "2025-12-31", "url": "https://arxiv.org/abs/2512.24965"}
{"aliases": [], "authors": ["Mingyu Ouyang", "Kevin Qinghong Lin", "Mike Zheng Shou", "Hwee Tou Ng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "518a9699a6c11e6f61b655f4016170c50bf5d949e477134521df1c0214320046", "submitted_at": "2026-01-07T13:48:12Z", "updated_at": "2026-01-07T13:48:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.03928v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.03928", "identifiers": {"arxiv": "2601.03928"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.03928"}], "name": "FocusUI: Efficient UI Grounding via Position-Preserving Visual Token Selection", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-01-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "FocusUI selects instruction-relevant screenshot patches while preserving their positional structure for GUI grounding. Its PosPad scheme replaces runs of dropped tokens with positional markers and is evaluated under reduced token budgets.", "topics": ["grounding", "efficiency", "screen-understanding"], "updated_at": "2026-01-07", "url": "https://arxiv.org/abs/2601.03928"}
{"aliases": [], "authors": ["Chen Chen", "Jiawei Shao", "Dakuan Lu", "Haoyi Hu", "Xiangcheng Liu", "Hantao Yao", "Wu Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d5ed7e72534fc55d028d4f25cb98b2373c88b622cf76396933d7a49d8ecf2e8", "submitted_at": "2026-01-14T14:27:28Z", "updated_at": "2026-01-14T14:27:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.09770v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.09770", "identifiers": {"arxiv": "2601.09770"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.09770"}], "name": "GUI-Eyes: Tool-Augmented Perception for Visual Grounding in GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Eyes trains GUI grounders to decide when and how to use visual tools such as crops and zooms. Its coarse-to-fine policy uses dense spatial rewards based on location proximity and region overlap, with evaluation on ScreenSpot-Pro.", "topics": ["grounding", "reinforcement-learning", "tool-integration", "screen-understanding"], "updated_at": "2026-01-14", "url": "https://arxiv.org/abs/2601.09770"}
{"aliases": [], "authors": ["Yurun Song", "Jiong Yin", "Rongjunchen Zhang", "Ian G. Harris"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "600f2a44695886fc527b04e1a8b1ad07cec90cf5c58dee834605d5133a760b1e", "submitted_at": "2026-01-14T03:52:57Z", "updated_at": "2026-01-14T03:52:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.11631v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.11631", "identifiers": {"arxiv": "2601.11631"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.11631"}], "name": "Compress to Focus: Efficient Coordinate Compression for Policy Optimization in Multi-Turn GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "CCPO couples GUI policy optimization with coordinate-guided compression of past visual observations. It aggregates targets across rollouts to focus retained context and uses distance-based advantages to train grounding and compression together.", "topics": ["grounding", "reinforcement-learning", "efficiency", "interaction-history"], "updated_at": "2026-01-14", "url": "https://arxiv.org/abs/2601.11631"}
{"aliases": [], "authors": ["Yanxi Wang", "Zhiling Zhang", "Wenbo Zhou", "Weiming Zhang", "Jie Zhang", "Qiannan Zhu", "Yu Shi", "Shuxin Zheng", "Jiyan He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ca912ad67a681e8e908a5a403e982f39b5080bee4df376eb0b598468f3bb504", "submitted_at": "2026-01-26T11:33:40Z", "updated_at": "2026-05-13T07:11:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.18842v3"}}, "environments": ["android", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.18842", "identifiers": {"arxiv": "2601.18842"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18842"}], "name": "GUIGuard-Bench: Toward a General Evaluation for Privacy-Preserving GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUIGuard-Bench, which annotates private regions in GUI trajectories with categories, risk levels, and relevance to the user’s task. It evaluates privacy recognition and the effect of screenshot protection on offline planning and task utility.", "topics": ["safety-security", "evaluation", "grounding"], "updated_at": "2026-05-13", "url": "https://arxiv.org/abs/2601.18842"}
{"aliases": [], "authors": ["Libo Sun", "Jiwen Zhang", "Siyuan Wang", "Zhongyu Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1930b4efca7ee38fc6cd1ceb283c14ed9b1388e7b758b951ba9469ed3067c3e0", "submitted_at": "2026-01-27T04:58:48Z", "updated_at": "2026-01-31T15:17:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.19199v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.19199", "identifiers": {"arxiv": "2601.19199"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.19199"}], "name": "MAGNET: Towards Adaptive GUI Agents with Memory-Driven Knowledge Evolution", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAGNET maintains separate memories for stable interface functions and task procedures across changing layouts and workflows. It updates frequently used knowledge and evaluates adaptation in AndroidWorld and offline distribution-shift settings.", "topics": ["memory", "grounding", "robustness", "evaluation"], "updated_at": "2026-01-31", "url": "https://arxiv.org/abs/2601.19199"}
{"aliases": [], "authors": ["Le Zhang", "Yixiong Xiao", "Xinjiang Lu", "Jingjia Cao", "Yusai Zhao", "Jingbo Zhou", "Lang An", "Zikan Feng", "Wanxiang Sha", "Yu Shi", "Congxi Xiao", "Jian Xiong", "Yankai Zhang", "Hua Wu", "Haifeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ad764cf86a2ecab81f3f730d6d26ea582544a355a90a8be55be220c98aabd7bf", "submitted_at": "2026-01-28T08:45:17Z", "updated_at": "2026-01-28T08:45:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.20380v1"}}, "environments": ["android", "desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.20380", "identifiers": {"arxiv": "2601.20380"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20380"}], "name": "OmegaUse: Building a General-Purpose GUI Agent for Autonomous Task Execution", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "OmegaUse combines exploratory and taxonomy-guided data synthesis with supervised fine-tuning followed by GRPO for GUI control. The report introduces OS-Nav for offline Android and Ubuntu evaluation alongside established grounding and action-prediction benchmarks.", "topics": ["data-generation", "reinforcement-learning", "grounding", "evaluation"], "updated_at": "2026-01-28", "url": "https://arxiv.org/abs/2601.20380"}
{"aliases": [], "authors": ["Ziwei Liu", "Borui Kang", "Hangjie Yuan", "Zixiang Zhao", "Wei Li", "Yifan Zhu", "Tao Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1083cf063ccf1781f2cdfe5e266b75b52079fd12ad16d05755577b880c010aec", "submitted_at": "2026-01-28T16:06:31Z", "updated_at": "2026-03-25T10:25:50Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2601.20732v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.20732", "identifiers": {"arxiv": "2601.20732"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20732"}], "name": "Continual GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Continual GUI Agents defines grounding under changing application domains and screen resolutions. Its GUI-AiF training framework uses separate point- and region-alignment rewards to adapt grounding as these distributions shift.", "topics": ["grounding", "reinforcement-learning", "robustness", "evaluation"], "updated_at": "2026-03-25", "url": "https://arxiv.org/abs/2601.20732"}
{"aliases": [], "authors": ["Qingni Wang", "Yue Fan", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9557fc608a7fd7667746e9b3e06bc699b1796cdd77fe5e975ba1bd5a64690c1", "submitted_at": "2026-02-02T18:22:45Z", "updated_at": "2026-02-03T08:31:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.02419v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.02419", "identifiers": {"arxiv": "2602.02419"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02419"}], "name": "SafeGround: Know When to Trust GUI Grounding Models via Uncertainty Calibration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SafeGround, which estimates GUI-grounding uncertainty from the spatial dispersion of sampled predictions and calibrates a decision threshold before testing. It studies risk-aware selection on ScreenSpot-Pro, with statistical control scoped to the threshold’s false-discovery rate.", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-02-03", "url": "https://arxiv.org/abs/2602.02419"}
{"aliases": [], "authors": ["Aiden Yiliu Li", "Xinyue Hao", "Shilong Liu", "Mengdi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5b730fa81936d6a1b2b0a8b74a522a22de39b977929165c44a5050e490c301f", "submitted_at": "2026-02-02T18:50:07Z", "updated_at": "2026-02-02T18:50:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02468v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.02468", "identifiers": {"arxiv": "2602.02468"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02468"}], "name": "Avenir-Web: Human-Experience-Imitating Multimodal Web Agents with Mixture of Grounding Experts", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Avenir-Web combines multiple grounding specialists with procedural experience, a task checklist and adaptive memory. It evaluates this planning and execution framework on live Online-Mind2Web tasks.", "topics": ["grounding", "planning", "memory", "web-navigation"], "updated_at": "2026-02-02", "url": "https://arxiv.org/abs/2602.02468"}
{"aliases": [], "authors": ["Xiaoce Wang", "Guibin Zhang", "Junzhe Li", "Jinzhe Tu", "Chun Li", "Ming Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49b0dbff31dcf8c966bebd367ae05984e46a7fd13b2cfb1ad5ef061332b43f00", "submitted_at": "2026-01-30T08:38:05Z", "updated_at": "2026-01-30T08:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.02548", "identifiers": {"arxiv": "2602.02548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02548"}], "name": "ToolTok: Tool Tokenization for Efficient and Generalizable GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "ToolTok represents GUI operations with learned tool tokens and frames interaction as progressive tool selection. Semantic anchoring and an easy-to-hard curriculum teach tool meanings before visual pathfinding, with evaluation across grounding scenarios.", "topics": ["tool-integration", "grounding", "data-generation"], "updated_at": "2026-01-30", "url": "https://arxiv.org/abs/2602.02548"}
{"aliases": [], "authors": ["Longhui Ma", "Di Zhao", "Siwei Wang", "Zhao Lv", "Miao Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2591e0dab6a84b5db2631ddc94a2c0b65b13a97d06d411ab6ad4e243fe3d3e1e", "submitted_at": "2026-02-06T03:27:55Z", "updated_at": "2026-02-06T03:27:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.06351v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.06351", "identifiers": {"arxiv": "2602.06351"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.06351"}], "name": "Trifuse: Enhancing Attention-Based GUI Grounding via Multimodal Fusion", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trifuse combines attention-based localization with OCR text cues and icon captions to ground GUI instructions. Its fusion procedure favors cross-modal agreement and sharp location peaks, and is evaluated without task-specific fine-tuning.", "topics": ["grounding", "screen-understanding", "training-free"], "updated_at": "2026-02-06", "url": "https://arxiv.org/abs/2602.06351"}
{"aliases": [], "authors": ["Zhongyin Zhao", "Yuan Liu", "Yikun Liu", "Haicheng Wang", "Le Tian", "Xiao Zhou", "Yangxiu You", "Zilin Yu", "Yang Yu", "Jie Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "852813305299178baeda3540cbad539dc30b840681e7bbdaed34dac1341dc0de", "submitted_at": "2026-02-06T05:14:11Z", "updated_at": "2026-02-06T05:14:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.06391v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.06391", "identifiers": {"arxiv": "2602.06391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.06391"}], "name": "POINTS-GUI-G: GUI-Grounding Journey", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies GUI-grounding development from a weakly grounded base model through data preparation, visual training, and verifiable rewards.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-02-06", "url": "https://arxiv.org/abs/2602.06391"}
{"aliases": [], "authors": ["Haodong Li", "Jingwei Wu", "Quan Sun", "Guopeng Li", "Juanxi Tian", "Ziyu Guo", "Donghao Zhou", "Tianfei Ren", "Huanyu Zhang", "Yanlin Lai", "Ruichuan An", "Hongbo Peng", "Yuhong Dai", "Jia Wang", "Ziyang Meng", "Zheng Ge", "Xiangyu Zhang", "Daxin Jiang", "Pheng-Ann Heng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43a01e0eb848a57a84767836d2c3c982bbf6094d22a3e70652dd42af52760bc5", "submitted_at": "2026-02-09T18:52:02Z", "updated_at": "2026-09-26T18:28:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.09007v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.09007", "identifiers": {"arxiv": "2602.09007"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09007"}], "name": "GUI-GenBench: Evaluating Image Generation Models as Interactive GUI Environments", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-02-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-GenBench tests image-generation models as action-conditioned GUI environments across five suites, including transitions, trajectories and grounding. Its VLM-based GUI-Score separates goal completion, interaction logic, consistency and visual properties.", "topics": ["evaluation", "world-models", "grounding", "screen-understanding"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2602.09007"}
{"aliases": [], "authors": ["Venus Team", "Changlong Gao", "Zhangxuan Gu", "Yulin Liu", "Xinyu Qiu", "Shuheng Shen", "Yue Wen", "Tianyu Xia", "Zhenyu Xu", "Zhengwen Zeng", "Beitong Zhou", "Xingran Zhou", "Weizhi Chen", "Sunhao Dai", "Jingya Dou", "Yichen Gong", "Yuan Guo", "Zhenlin Guo", "Feng Li", "Qian Li", "Jinzhen Lin", "Yuqi Zhou", "Linchao Zhu", "Liang Chen", "Zhenyu Guo", "Changhua Meng", "Weiqiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efc4ad8f13d55939abcc783bdd652f276c3ac0164ea35232a3fb9366642d4558", "submitted_at": "2026-02-09T18:43:40Z", "updated_at": "2026-02-24T15:17:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.09082v2"}}, "environments": ["cross-platform", "web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.09082", "identifiers": {"arxiv": "2602.09082"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09082"}], "name": "UI-Venus-1.5 Technical Report", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a GUI-agent family combining broad intermediate training, full-trajectory online reinforcement learning, and merging of specialized models.", "topics": ["reinforcement-learning", "grounding", "mobile-navigation"], "updated_at": "2026-02-24", "url": "https://arxiv.org/abs/2602.09082"}
{"aliases": [], "authors": ["Xingyi Zhang", "Yulei Ye", "Kaifeng Huang", "Wenhao Li", "Xiangfeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86a47f084b23cc70e754808f6404bed0bbd500f78fcac4ccc104f95052a49868", "submitted_at": "2026-02-11T12:54:53Z", "updated_at": "2026-02-11T12:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.10814v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.10814", "identifiers": {"arxiv": "2602.10814"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.10814"}], "name": "See, Plan, Snap: Evaluating Multimodal GUI Agents in Scratch", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScratchWorld evaluates program construction through the Scratch interface using fine-grained drag-and-drop actions and a complementary semantic-API mode. Runtime tests assess the resulting programs, allowing analysis of the gap between program reasoning and GUI manipulation.", "topics": ["evaluation", "grounding", "automation"], "updated_at": "2026-02-11", "url": "https://arxiv.org/abs/2602.10814"}
{"aliases": [], "authors": ["Yibo Wang", "Guangda Huzhang", "Yuwei Hu", "Yu Xia", "Shiyin Lu", "Qing-Guo Chen", "Zhao Xu", "Weihua Luo", "Kaifu Zhang", "Lijun Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "573d859133ec9986af92e1943d0ec913d914acf4e54a030a4b103181ba505ae0", "submitted_at": "2026-02-14T07:44:47Z", "updated_at": "2026-02-14T07:44:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13653v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.13653", "identifiers": {"arxiv": "2602.13653"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.13653"}], "name": "Building Autonomous GUI Navigation via Agentic-Q Estimation and Step-Wise Policy Optimization", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "This framework learns an agentic-Q model to estimate individual actions’ contributions to GUI task completion. It then optimizes a policy with step-level samples from its own trajectories, decoupling policy updates from environment interaction.", "topics": ["reinforcement-learning", "reward-models", "grounding"], "updated_at": "2026-02-14", "url": "https://arxiv.org/abs/2602.13653"}
{"aliases": [], "authors": ["A. Said Gurbuz", "Sunghwan Hong", "Ahmed Nassar", "Marc Pollefeys", "Peter Staar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7c6e23f704562c4ae4db9e0b4f8f2a8dde1318efc08876ee57f5967b178257e", "submitted_at": "2026-02-15T19:00:02Z", "updated_at": "2026-05-01T17:32:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.14276v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.14276", "identifiers": {"arxiv": "2602.14276"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14276"}], "name": "ScreenParse: Moving Beyond Sparse Grounding with Complete Screen Parsing Supervision", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScreenParse supplies dense annotations of UI elements, types and text in web screenshots through an automated rendering and filtering pipeline. ScreenVLM learns to output structured screen markup, and the study evaluates dense parsing and transfer to GUI grounding.", "topics": ["screen-understanding", "data-generation", "grounding", "evaluation"], "updated_at": "2026-05-01", "url": "https://arxiv.org/abs/2602.14276"}
{"aliases": [], "authors": ["Sicheng Fan", "Rui Wan", "Yifei Leng", "Gaoning Liang", "Li Ling", "Yanyi Shang", "Dehan Kong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db52246d9f5f305046e6dd14f2ef15440c371678e1db0d05282c84a3667c6f10", "submitted_at": "2026-03-05T15:37:34Z", "updated_at": "2026-04-14T08:41:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.05295v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2603.05295", "identifiers": {"arxiv": "2603.05295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.05295"}], "name": "WebChain: A Large-Scale Human-Annotated Dataset of Real-World Web Interaction Traces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebChain aligns human web-interaction trajectories with visual, structural and action data for agent training. It also introduces WebChainBench and a mid-training approach that separates spatial grounding from planning.", "topics": ["data-generation", "evaluation", "grounding", "planning"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2603.05295"}
{"aliases": [], "authors": ["Xunzhuo Liu", "Bowei He", "Xue Liu", "Andy Luo", "Haichen Zhang", "Huamin Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "274deff6cacfb5f1b7b620a86abe7ef6496b673d929de39376a9ad42dd3c4c02", "submitted_at": "2026-03-13T09:21:25Z", "updated_at": "2026-03-13T09:21:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12823v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.12823", "identifiers": {"arxiv": "2603.12823"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.12823"}], "name": "Adaptive Vision-Language Model Routing for Computer Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-03-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adaptive VLM Routing estimates GUI-action difficulty and small-model confidence to select a model under a predicted reliability threshold. Grounding and agent-routing benchmarks assess the cost–accuracy trade-off, with prior interaction context used for warm agents.", "topics": ["efficiency", "grounding", "memory", "evaluation"], "updated_at": "2026-03-13", "url": "https://arxiv.org/abs/2603.12823"}
{"aliases": [], "authors": ["Ziwei Liu", "Tao Feng", "Borui Kang", "Yanbing Yang", "Jun Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4507a652e2b3c3857289763442fffd8db21726b97fb7b6304ad61a2139ef5af", "submitted_at": "2026-03-15T15:47:47Z", "updated_at": "2026-03-15T15:47:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.14448v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.14448", "identifiers": {"arxiv": "2603.14448"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.14448"}], "name": "Zoom to Essence: Trainless GUI Grounding by Inferring upon Interface Elements", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents ZoomUI, a training-free grounding method that expresses instructions as visual element descriptions and uses model attention to zoom progressively into relevant interface regions. Evaluation examines instruction-to-element localization on GUI benchmarks.", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-03-15", "url": "https://arxiv.org/abs/2603.14448"}
{"aliases": [], "authors": ["Xunzhuo Liu", "Bowei He", "Xue Liu", "Andy Luo", "Haichen Zhang", "Huamin Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c7cb0e4784c2a58feeca27a1687d1f73880d9713c16387f0bbff9ed34c427c7", "submitted_at": "2026-03-16T01:31:32Z", "updated_at": "2026-03-16T01:31:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.14707v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.14707", "identifiers": {"arxiv": "2603.14707"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.14707"}], "name": "Visual Confused Deputy: Exploiting and Defending Perception Failures in Computer-Using Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Formalizes a visual confused-deputy failure in which a computer agent acts on a misperceived UI object. A separate guardrail checks both the visual target and the agent's action reasoning against deployment-specific knowledge before permitting execution.", "topics": ["safety-security", "grounding", "robustness"], "updated_at": "2026-03-16", "url": "https://arxiv.org/abs/2603.14707"}
{"aliases": [], "authors": ["Siqi Pei", "Liang Tang", "Tiaonan Duan", "Long Chen", "Shuxian Li", "Kaer Huang", "Yanzhe Jing", "Yiqiang Yan", "Bo Zhang", "Chenghao Jiang", "Borui Zhang", "Jiwen Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2101a32746d6c668bff79461983b2e12138959f49ab3cfcb542ecbd629317f1d", "submitted_at": "2026-03-18T07:26:18Z", "updated_at": "2026-03-18T07:26:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.17441v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.17441", "identifiers": {"arxiv": "2603.17441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.17441"}], "name": "AdaZoom-GUI: Adaptive Zoom-based GUI Grounding with Instruction Refinement", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines instruction rewriting with conditional second-stage zooming for small GUI targets. A grounding dataset and GRPO training support joint prediction of click coordinates and element bounding boxes.", "topics": ["grounding", "reinforcement-learning", "screen-understanding"], "updated_at": "2026-03-18", "url": "https://arxiv.org/abs/2603.17441"}
{"aliases": [], "authors": ["Rozain Shakeel", "Abdul Rahman Mohammad Ali", "Muneeb Mushtaq", "Tausifa Jan Saleem", "Tajamul Ashraf"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfa6cd9aead44150fdc997654ded7253b2890e936b94e44be2fb33642b24d302", "submitted_at": "2026-03-20T14:43:53Z", "updated_at": "2026-03-20T14:43:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19993v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2603.19993", "identifiers": {"arxiv": "2603.19993"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.19993"}], "name": "MedSPOT: A Workflow-Aware Sequential Grounding Benchmark for Clinical GUI", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-03-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Benchmarks sequential target localization in clinical software workflows, emphasizing how early mistakes affect subsequent interaction steps.", "topics": ["grounding", "evaluation"], "updated_at": "2026-03-20", "url": "https://arxiv.org/abs/2603.19993"}
{"aliases": [], "authors": ["Xiangru Jian", "Shravan Nayak", "Kevin Qinghong Lin", "Aarash Feizi", "Kaixin Li", "Patrice Bechard", "Spandana Gella", "Sai Rajeswar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15237fda2f3799314009635017487749dbedc9b1a0ffdd6b56a8d28eaf01481d", "submitted_at": "2026-03-25T15:52:56Z", "updated_at": "2026-03-25T15:52:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24440v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.24440", "identifiers": {"arxiv": "2603.24440"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.24440"}], "name": "CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "CUA-Suite combines continuous expert desktop demonstrations in VideoCUA with the UI-Vision evaluation benchmark and GroundCUA grounding data. Screen recordings, cursor traces and layered annotations support study of temporal behavior and professional software workflows.", "topics": ["data-generation", "video", "evaluation", "grounding", "workflow"], "updated_at": "2026-03-25", "url": "https://arxiv.org/abs/2603.24440"}
{"aliases": [], "authors": ["Shrinidhi Kumbhar", "Haofu Liao", "Srikar Appalaraju", "Kunwar Yashraj Singh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0838afd6777af3e3995fa3bf9645102f98962cccd81c1db438a007b4cd33c442", "submitted_at": "2026-03-27T09:34:07Z", "updated_at": "2026-03-27T09:34:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.26211v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.26211", "identifiers": {"arxiv": "2603.26211"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.26211"}], "name": "Towards GUI Agents: Vision-Language Diffusion Models for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts a discrete vision-language diffusion model to single-turn GUI action and bounding-box prediction, using a hybrid masking schedule for coordinate geometry. Experiments across web, desktop and mobile datasets examine grounding accuracy and inference tradeoffs.", "topics": ["grounding", "action-prediction", "efficiency"], "updated_at": "2026-03-27", "url": "https://arxiv.org/abs/2603.26211"}
{"aliases": [], "authors": ["Rui Xie", "Zhi Gao", "Chenrui Shi", "Zirui Shang", "Lu Chen", "Qing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b388086aeb8ac47e386099e5cd68cb89bb03978ee7d5897950181035d0abb723", "submitted_at": "2026-03-27T10:33:08Z", "updated_at": "2026-06-30T04:47:42Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.26266v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.26266", "identifiers": {"arxiv": "2603.26266"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.26266"}], "name": "GUIDE: Resolving Domain Bias in GUI Agents through Real-Time Web Video Retrieval and Plug-and-Play Annotation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-03-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUIDE retrieves relevant software tutorials through subtitle-based video search, then infers planning and grounding knowledge from consecutive annotated keyframes. It inserts that knowledge into existing GUI-agent modules without parameter updates and evaluates the plug-in on OSWorld.", "topics": ["video", "planning", "grounding", "training-free"], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2603.26266"}
{"aliases": [], "authors": ["Christopher Clark", "Yue Yang", "Jae Sung Park", "Zixian Ma", "Jieyu Zhang", "Rohun Tripathi", "Mohammadreza Salehi", "Sangho Lee", "Taira Anderson", "Winson Han", "Ranjay Krishna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fe857c9f7e7819a9c2f99448f77bbeae15ae4e2ac8ad308000f3d217b0177fc", "submitted_at": "2026-03-30T06:15:06Z", "updated_at": "2026-03-30T06:15:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.28069v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.28069", "identifiers": {"arxiv": "2603.28069"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.28069"}], "name": "MolmoPoint: Better Pointing for VLMs with Grounding Tokens", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent general VLM pointing method with an explicit GUI-grounding experiment; no autonomous GUI-task execution is established.", "published_at": "2026-03-30", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "MolmoPoint replaces textual coordinate generation with pointing tokens that select visual regions and refine them through subpatches. The general image/video method includes GUI-pointing evaluation on ScreenSpotPro alongside other visual grounding tasks.", "topics": ["grounding", "screen-understanding", "video", "efficiency"], "updated_at": "2026-03-30", "url": "https://arxiv.org/abs/2603.28069"}
{"aliases": [], "authors": ["Songze Li", "Xiaoke Guo", "Tianqi Liu", "Biao Yi", "Zhaoyan Gong", "Zhiqiang Liu", "Huajun Chen", "Wen Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b99643e43c5e20af088d6fa62152e03256c947af76a6b3c0abfc6f74a0f7d648", "submitted_at": "2026-04-08T12:12:09Z", "updated_at": "2026-10-02T06:30:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.06995v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.06995", "identifiers": {"arxiv": "2604.06995"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06995"}], "name": "What's Missing in Screen-to-Action? Towards a UI-in-the-Loop Paradigm for Multimodal GUI Reasoning", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UILoop, which inserts explicit UI-element localization, function and usage reasoning between screen observation and action. UI Comprehension-Bench evaluates these element-level capabilities with three metrics.", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2604.06995"}
{"aliases": [], "authors": ["Wenkui Yang", "Chao Jin", "Haisu Zhu", "Weilin Luo", "Derek Yuen", "Kun Shao", "Junxian Duan", "Huaibo Huang", "Jie Cao", "Ran He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5427eedff0e97f9c73101403754c38e8928ad71abab7b7eff5f136a30ee2205", "submitted_at": "2026-04-09T05:32:34Z", "updated_at": "2026-07-08T02:39:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.07831v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.07831", "identifiers": {"arxiv": "2604.07831"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.07831"}], "name": "Are GUI Agents Focused Enough? Automated Distraction via Semantic-level UI Element Injection", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study tests black-box GUI-grounding failures caused by strategically added, innocuous-looking UI elements. Its iterative overlay search is compared with random injection and repeated independent grounding trials to distinguish persistent misdirection from incidental clutter.", "topics": ["safety-security", "grounding", "robustness", "evaluation"], "updated_at": "2026-07-08", "url": "https://arxiv.org/abs/2604.07831"}
{"aliases": [], "authors": ["Wee Joe Tan", "Zi Rui Lucas Lim", "Shashank Durgad", "Karim Obegi", "Aiden Yiliu Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "445e520a60e0e3cc0c3c519e21bfa8eebfa33d4fe4320c437e67b94a02d84504", "submitted_at": "2026-02-25T18:59:42Z", "updated_at": "2026-04-14T20:54:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.09581v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.09581", "identifiers": {"arxiv": "2604.09581"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09581"}], "name": "Avenir-UX: Automated UX Evaluation via Simulated Human Web Interaction with GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Avenir-UX combines grounded web interaction with simulated user profiles and structured usability measures to produce evaluations of website journeys. The paper describes its architecture and use of SUS, step-wise ease questions and think-aloud traces for automated UX reporting.", "topics": ["automation", "evaluation", "ui-design", "grounding"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2604.09581"}
{"aliases": [], "authors": ["Himangi Mittal", "Gaurav Mittal", "Nelson Daniel Troncoso", "Yu Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa89e3a09c0ce8b09bdb932f3f00bd2dccf8af4da11a727e675ddf8b2c578146", "submitted_at": "2026-04-14T17:55:46Z", "updated_at": "2026-05-28T06:47:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.13019v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.13019", "identifiers": {"arxiv": "2604.13019"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.13019"}], "name": "PrecisionCUA: Iterative Visual Refinement for Pixel-Precise Cursor Grounding in Code Editors", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies iterative cursor grounding in dense code-editor interfaces, using visual feedback from earlier attempts to correct displacement. PrecisionCUA evaluates this closed-loop process across multiple model families on coding-interface tasks.", "topics": ["grounding", "robustness", "screen-understanding"], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2604.13019"}
{"aliases": [], "authors": ["Fei Tang", "Bofan Chen", "Zhengxi Lu", "Tongbo Chen", "Songqin Nong", "Tao Jiang", "Wenhao Xu", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68709f498ebd5d475593ca8ac5bd543a695086d4dffadc113fad908495c234d9", "submitted_at": "2026-04-15T17:32:28Z", "updated_at": "2026-04-15T17:32:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14113v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.14113", "identifiers": {"arxiv": "2604.14113"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14113"}], "name": "UI-Zoomer: Uncertainty-Driven Adaptive Zoom-In for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UI-Zoomer, a training-free grounding method that uses candidate agreement and token confidence to decide when to crop and retry. Prediction variance determines a separate crop radius for each uncertain interface target.", "topics": ["grounding", "training-free", "efficiency"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2604.14113"}
{"aliases": [], "authors": ["Yangyue Wang", "Harshvardhan Sikka", "Yash Mathur", "Tony Zhou", "Jinu Nyachhyon", "Pranav Guruprasad"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5dc36ff0c9c4bd39de360824967e53c18fe57562cb8da1462d3e63c0dd348e71", "submitted_at": "2026-04-15T16:39:22Z", "updated_at": "2026-07-01T10:12:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.14262v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.14262", "identifiers": {"arxiv": "2604.14262"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14262"}], "name": "GUI-Perturbed: Domain Randomization Reveals Systematic Brittleness in GUI Grounding Models", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-Perturbed, which varies screenshot appearance and instruction wording independently to diagnose grounding failures. Experiments on three related model architectures examine relational instructions, browser zoom and augmentation-based fine-tuning.", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2604.14262"}
{"aliases": [], "authors": ["Pan Hao", "Rishi Selvakumaran", "Jacob Sun", "Qianwen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd3b5e5b9a607023c82150673255eaf2f90c22db86d3c049b2a2fd536aaef17f", "submitted_at": "2026-04-16T06:22:48Z", "updated_at": "2026-04-16T06:22:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14668v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.14668", "identifiers": {"arxiv": "2604.14668"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14668"}], "name": "Beyond Chat and Clicks: GUI Agents for In-Situ Assistance via Live Interface Transformation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "DOMSteer translates help requests and live page context into reversible DOM changes such as control highlighting, contextual tooltips and layout reorganization. The work evaluates in-place GUI assistance with interface tests, use cases and a comparative user study.", "topics": ["human-in-the-loop", "ui-design", "grounding", "evaluation"], "updated_at": "2026-04-16", "url": "https://arxiv.org/abs/2604.14668"}
{"aliases": [], "authors": ["Chao Jin", "Wenkui Yang", "Hao Sun", "Yuqi Liao", "Qianyi Jiang", "Kai Zhou", "Jie Cao", "Ran He", "Huaibo Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b102c9363b1be77273c7fab6f7f5cde74157c646671c93b9503770fe60dd70c", "submitted_at": "2026-04-19T06:55:59Z", "updated_at": "2026-04-19T06:55:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17284v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.17284", "identifiers": {"arxiv": "2604.17284"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17284"}], "name": "HalluClear: Diagnosing, Evaluating and Mitigating Hallucinations in GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces HalluClear with a GUI hallucination taxonomy, expert-calibrated evaluation workflow and structured-reasoning post-training scheme. It evaluates how these components affect hallucinations, grounding and action fidelity across GUI agents.", "topics": ["robustness", "grounding", "evaluation"], "updated_at": "2026-04-19", "url": "https://arxiv.org/abs/2604.17284"}
{"aliases": [], "authors": ["Wenkai Wang", "Xiyun Li", "Hongcan Guo", "Wenhao Yu", "Tianqing Fang", "Haitao Mi", "Dong Yu", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e15661746ab05f2894e23d28b4675e13919bbcd71407fbc33260d84f330ea5d", "submitted_at": "2026-04-23T04:23:31Z", "updated_at": "2026-04-23T04:23:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.21268v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.21268", "identifiers": {"arxiv": "2604.21268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.21268"}], "name": "Measure Twice, Click Once: Co-evolving Proposer and Visual Critic via Reinforcement Learning for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains a GUI grounding proposer together with a visual critic that evaluates candidate locations rendered on screenshots. An adaptive reinforcement-learning schedule balances their development, using proposal diversity and critic feedback during training.", "topics": ["grounding", "reinforcement-learning", "self-reflection"], "updated_at": "2026-04-23", "url": "https://arxiv.org/abs/2604.21268"}
{"aliases": [], "authors": ["Hongxin Li", "Yuntao Chen", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fe6a322737dd58339673bb1c956345ff585d79df6185ce1f9eaa452b37a1e01", "submitted_at": "2026-04-27T01:29:02Z", "updated_at": "2026-04-27T01:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.23941v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2604.23941", "identifiers": {"arxiv": "2604.23941"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.23941"}], "name": "GoClick: Lightweight Element Grounding Model for Autonomous GUI Interaction", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-04-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies a compact encoder-decoder grounding model and progressive training-data refinement for resource-constrained GUI interaction.", "topics": ["grounding", "efficiency"], "updated_at": "2026-04-27", "url": "https://arxiv.org/abs/2604.23941"}
{"aliases": [], "authors": ["Hongxin Li", "Xiping Wang", "Jingran Su", "Zheng Ju", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65325912365effe8adbc39301a378d7526c60d1185691f3a89cb8acd4aa53c3f", "submitted_at": "2026-04-27T13:06:27Z", "updated_at": "2026-04-27T13:06:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24441v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.24441", "identifiers": {"arxiv": "2604.24441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.24441"}], "name": "AutoGUI-v2: A Comprehensive Multi-Modal GUI Functionality Understanding Benchmark", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "AutoGUI-v2 evaluates interface functionality, grounding and predicted interaction outcomes using hierarchical regions derived through VLM-human annotation. Its tasks span multiple operating systems and expose differences between functional localization, captioning and transition-logic understanding.", "topics": ["screen-understanding", "grounding", "state-tracking", "evaluation", "data-generation"], "updated_at": "2026-04-27", "url": "https://arxiv.org/abs/2604.24441"}
{"aliases": [], "authors": ["Michito Takeshita", "Takuro Kawada", "Takumi Ohashi", "Shunsuke Kitada", "Hitoshi Iyatomi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87ef518590e29328ce752791df974c9e2215eac656fcf43fed97dd3e847f3cf7", "submitted_at": "2026-05-01T10:16:00Z", "updated_at": "2026-05-01T10:16:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.00551v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.00551", "identifiers": {"arxiv": "2605.00551"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.00551"}], "name": "A11y-Compressor: A Framework for Enhancing the Efficiency of GUI Agent Observations through Visual Context Reconstruction and Redundancy Reduction", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "A11y-Compressor transforms linearized accessibility trees through modal detection, redundancy reduction and semantic structuring. Its OSWorld evaluation measures both observation-token reduction and task success.", "topics": ["screen-understanding", "efficiency", "grounding", "evaluation"], "updated_at": "2026-05-01", "url": "https://arxiv.org/abs/2605.00551"}
{"aliases": [], "authors": ["Ruilin Yao", "Shegnwu Xiong", "Tianyu Zou", "Shili Xiong", "Yi Rong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dc8c2f5a5a97bc372a1bf5ab3fe0537e53a681f808d45877f55cdef6556858c", "submitted_at": "2026-05-04T14:18:46Z", "updated_at": "2026-05-04T14:18:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.02630v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.02630", "identifiers": {"arxiv": "2605.02630"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.02630"}], "name": "AutoFocus: Uncertainty-Aware Active Visual Search for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents AutoFocus, a training-free grounding method that turns sampled coordinates and token uncertainty into spatial region proposals. Shape-aware crops and visual comparison select a target while retaining surrounding interface context.", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-05-04", "url": "https://arxiv.org/abs/2605.02630"}
{"aliases": [], "authors": ["Xiaoming Ren", "Ru Zhen", "Chao Li", "Yang Song", "Qiuxia Hou", "Yanhao Zhang", "Peng Liu", "Qi Qi", "Quanlong Zheng", "Qi Wu", "Zhenyi Liao", "Binqiang Pan", "Haobo Ji", "Haonan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ece93c00c5556dfe50ddd0ba3f52b725f34d1bee18ea3565951bb78343d8ed0", "submitted_at": "2026-05-07T06:58:34Z", "updated_at": "2026-05-21T06:55:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.05765v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:10.101459+00:00", "role": "primary", "sha256": "80375cd810dcda2e9ccca2c6fbee44730de8449972036c548bde312566b20673", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2605.05765v2"}], "id": "arxiv:2605.05765", "identifiers": {"arxiv": "2605.05765"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:51.952206+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.05765"}], "name": "X-OmniClaw Technical Report: A Unified Mobile Agent for Multimodal Understanding and Interaction", "notes": "The report presents demonstrations. Its behavior-cloning component captures runtime skills; it does not establish policy-weight imitation training, and literal trace replay is described as future work. No general benchmark result or release completeness is inferred. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (2 Framework; 3 Omni Perception; 4 Omni Memory; 5.1 Omni Action in App Ecosystem; 5.2 Trajectory Cloned Execution); not an entire-paper reading or reproduction claim.", "published_at": "2026-05-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes an Android agent architecture combining multimodal perception, working and personal memory, hybrid XML/visual grounding, and reusable skills captured from user interactions.", "topics": ["mobile-navigation", "memory", "grounding", "workflow"], "updated_at": "2026-05-21", "url": "https://arxiv.org/abs/2605.05765"}
{"aliases": [], "authors": ["Miaosen Zhang", "Xiaohan Zhao", "Zhihong Tan", "Zhou Huoshen", "Yijia Fan", "Yifan Yang", "Kai Qiu", "Bei Liu", "Justin Wagle", "Chenzhong Yin", "Mingxi Cheng", "Ji Li", "Qi Dai", "Chong Luo", "Xu Yang", "Xin Geng", "Baining Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1fb57ba3cd858513d1da58d64a7c85388f12e331c9ea9f89204be5f43e6fec38", "submitted_at": "2026-05-12T17:59:58Z", "updated_at": "2026-05-12T17:59:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12501v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.12501", "identifiers": {"arxiv": "2605.12501"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12501"}], "name": "Covering Human Action Space for Computer Use: Data Synthesis and Benchmark", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CUActSpot to evaluate clicks, drags, drawing and other screen interactions across GUI, text, tables, canvases and natural images. A renderer-based pipeline records geometry and screenshots, then generates matching instructions and action traces for training a grounding model.", "topics": ["grounding", "data-generation", "evaluation", "action-prediction"], "updated_at": "2026-05-12", "url": "https://arxiv.org/abs/2605.12501"}
{"aliases": [], "authors": ["Jiaping Lin", "Fei Shen", "Junzhe Li", "Ping Nie", "Fei Yu", "Ming Li", "Haizhou Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb60b76fca3a8825030b1a31a7b97ff31e06f30ac3852f243efe440f965e5578", "submitted_at": "2026-05-10T07:04:07Z", "updated_at": "2026-05-10T07:04:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12549v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.12549", "identifiers": {"arxiv": "2605.12549"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12549"}], "name": "What Happens Before Decoding? Prefill Determines GUI Grounding in VLMs", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Re-Prefill adds an attention-guided second prefill pass that revisits candidate screen regions before coordinate generation. The study evaluates this training-free intervention across GUI-grounding models and benchmarks.", "topics": ["grounding", "training-free", "screen-understanding", "evaluation"], "updated_at": "2026-05-10", "url": "https://arxiv.org/abs/2605.12549"}
{"aliases": [], "authors": ["Yichao Liu", "Huawen Shen", "Liu Yu", "Shiyu Liu", "Zeyu Chen", "Yu Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd85c0d0a8b6bc5c687bc301b1b84d6a019c5713012890e0a23c08ae2d83a1e6", "submitted_at": "2026-05-15T02:27:41Z", "updated_at": "2026-05-15T02:27:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.15542v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.15542", "identifiers": {"arxiv": "2605.15542"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.15542"}], "name": "DRS-GUI: Dynamic Region Search for Training-Free GUI Grounding", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "DRS-GUI searches screenshot regions through focus, shift and scatter operations selected by a tree-search planner. A region-quality reward guides this training-free perceptual search before final GUI grounding.", "topics": ["grounding", "training-free", "planning", "screen-understanding"], "updated_at": "2026-05-15", "url": "https://arxiv.org/abs/2605.15542"}
{"aliases": [], "authors": ["Haoren Zhao", "Tianyi Chen", "Zhen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "671f33dd87a7ebf9233494ef6fa0efd9c9772a3951f3c218bf4de822825f7af2", "submitted_at": "2026-05-13T02:48:52Z", "updated_at": "2026-05-13T02:48:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16402v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.16402", "identifiers": {"arxiv": "2605.16402"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.16402"}], "name": "WinDeskGround: A Benchmark for Robust GUI Grounding in Complex Multi-Window Desktop Environments", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "WinDeskGround synthesizes multi-window desktop layouts with controlled occlusion, density and semantic similarity to test GUI grounding robustness. Its instruction-target benchmark measures how localization changes when interfaces depart from simple single-layer scenes.", "topics": ["grounding", "evaluation", "robustness", "data-generation"], "updated_at": "2026-05-13", "url": "https://arxiv.org/abs/2605.16402"}
{"aliases": [], "authors": ["Shilong Jin", "Lanjun Wang", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52c1bf4c614ac05a55f2c2c66d3ac4eeb1bf31743814a45eb9f0ce064b6baaab", "submitted_at": "2026-05-16T08:51:57Z", "updated_at": "2026-05-16T08:51:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16883v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.16883", "identifiers": {"arxiv": "2605.16883"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.16883"}], "name": "SE-GA: Memory-Augmented Self-Evolution for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "SE-GA retrieves hierarchical memories during inference and uses the collected experience in a separate policy-improvement pipeline. Its evaluation includes offline grounding/action tasks and online Android interaction, which measure different capabilities.", "topics": ["memory", "interaction-history", "mobile-navigation", "grounding", "evaluation"], "updated_at": "2026-05-16", "url": "https://arxiv.org/abs/2605.16883"}
{"aliases": [], "authors": ["Athar Parvez", "Muhammad Jawad Mufti", "Muqaddas Gull", "Omar Hammad"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff95e0b071508592190ba236741c1c01200dd38fdc74c31b2cc4d1bf8143ae9c", "submitted_at": "2026-05-17T21:14:32Z", "updated_at": "2026-05-17T21:14:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17656v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.17656", "identifiers": {"arxiv": "2605.17656"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.17656"}], "name": "MUIAnno: An Expert-Annotated Dataset and Evaluation Benchmark for Mobile UI Understanding", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-05-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "MUIAnno provides expert labels for elements in manually explored mobile-app screens, with a web annotation tool that emits structured JSON. It also supplies UI-element detection baselines for the dataset.", "topics": ["screen-understanding", "grounding", "evaluation", "data-generation"], "updated_at": "2026-05-17", "url": "https://arxiv.org/abs/2605.17656"}
{"aliases": [], "authors": ["Jingjing Liu", "Ziye Huang", "Zihao Cheng", "Zeming Liu", "Jiahong Wu", "Yuhang Guo", "Kehai Chen", "Yunhong Wang", "Haifeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "017909a697c6c661d3486d960042f4d691f70a660a60083e70512de34690cb51", "submitted_at": "2026-05-18T08:36:04Z", "updated_at": "2026-05-18T08:36:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.18048v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.18048", "identifiers": {"arxiv": "2605.18048"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.18048"}], "name": "DocOS: Towards Proactive Document-Guided Actions in GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "DocOS benchmarks agents that search online documentation and translate procedural instructions into GUI actions in interactive web environments. It separates failures to retrieve useful procedures from failures to ground retrieved instructions into executable actions.", "topics": ["web-navigation", "grounding", "evaluation", "planning"], "updated_at": "2026-05-18", "url": "https://arxiv.org/abs/2605.18048"}
{"aliases": [], "authors": ["Nathan Bout", "Maxime Langevin", "Ronan Riochet"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27ffcb10d966946de9fa2353c4b96798e5f090bdfdc5610f5e8a5c6233cf70ea", "submitted_at": "2026-06-04T15:57:29Z", "updated_at": "2026-06-04T15:57:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06322v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "id": "arxiv:2606.06322", "identifiers": {"arxiv": "2606.06322"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.06322"}], "name": "DragOn: A Benchmark and Dataset for Drag-Based GUI Interactions", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "DragOn supplies drag-grounding training data and a held-out benchmark for highlighting text, selecting cells, resizing elements and moving sliders. It evaluates existing vision-language models and a model fine-tuned on the drag tasks.", "topics": ["grounding", "data-generation", "evaluation", "screen-understanding"], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2606.06322"}
{"aliases": [], "authors": ["Radeen Mostafa", "Sawradip Saha"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d29dec745923e652654521f65a32f919a99d4577992006efc1a5e469bf775d81", "submitted_at": "2026-06-08T12:18:22Z", "updated_at": "2026-06-08T12:18:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09399v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.09399", "identifiers": {"arxiv": "2606.09399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09399"}], "name": "RunAgent SuperBrowser: A Theory of Autonomous Web Navigation Grounded in Human Browsing Behaviour", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "SUPERBROWSER combines screenshot-based target proposals with separate orchestration, planning and action roles. A bounded ledger and context-eviction process retain recent actions, task facts and checkpoints during browser execution.", "topics": ["web-navigation", "grounding", "planning", "memory", "efficiency"], "updated_at": "2026-06-08", "url": "https://arxiv.org/abs/2606.09399"}
{"aliases": [], "authors": ["Can Lin", "Tao Feng", "Hangjie Yuan", "Dan Zhang", "Yifan Zhu", "Zhonghong Ou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7523d3901755d8feeecbaf2f02dcc00fd0a40102e1498d49b9f789b0a7c02d75", "submitted_at": "2026-06-09T07:52:10Z", "updated_at": "2026-07-06T06:23:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.10522v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.10522", "identifiers": {"arxiv": "2606.10522"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.10522"}], "name": "GUI-AC: Enhancing Continual Learning in GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-AC uses grounding certainty to adjust advantage weighting and policy-clipping bounds during continual reinforcement fine-tuning. Its mechanisms aim to suppress noisy confidence updates while retaining exploration under shifting GUI domains and resolutions.", "topics": ["reinforcement-learning", "grounding", "robustness"], "updated_at": "2026-07-06", "url": "https://arxiv.org/abs/2606.10522"}
{"aliases": [], "authors": ["Jaewoo Lee", "Zaid Khan", "Archiki Prasad", "Justin Chih-Yao Chen", "Supriyo Chakraborty", "Kartik Balasubramaniam", "Sambit Sahu", "Elias Stengel-Eskin", "Hyunji Lee", "Mohit Bansal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8ed399c8b489a80c5313a00232d2ecca0fbc8228b1f56745c02d4fedb6f6908", "submitted_at": "2026-06-09T16:39:10Z", "updated_at": "2026-06-09T16:39:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.11078v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.11078", "identifiers": {"arxiv": "2606.11078"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.11078"}], "name": "A History-Aware Visually Grounded Critic for Computer Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "HiViG combines a compact history of completed actions with a visually grounded critic that checks proposed coordinates before execution. Its evaluations separate history-related planning failures from errors caused by targeting the wrong screen element.", "topics": ["interaction-history", "memory", "grounding", "reward-models", "long-horizon", "evaluation"], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2606.11078"}
{"aliases": [], "authors": ["Seoyoung Choi", "Minseok Ko", "Hyunseok Lee", "Kunwoong Kim", "Woomin Song", "Chanseok Jeon", "Jinwoo Shin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1fb87da27071ab21519599b7d16359a38118aa358f620de94e66b15cad040828", "submitted_at": "2026-06-12T04:39:01Z", "updated_at": "2026-06-12T04:39:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14106v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.14106", "identifiers": {"arxiv": "2606.14106"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.14106"}], "name": "Naive Visual Memory is Not Enough: A Failure-Mode Study of GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study finds that full-image visual memory can reduce some state errors while worsening action-level failures. Its AGMem alternative stores local crops associated with successful actions or recovery and evaluates their use in desktop GUI tasks.", "topics": ["memory", "interaction-history", "grounding", "robustness", "evaluation"], "updated_at": "2026-06-12", "url": "https://arxiv.org/abs/2606.14106"}
{"aliases": [], "authors": ["Xinyu Qiu", "Yunzhu Zhang", "Heng Jia", "Shuheng Shen", "Changhua Meng", "Linchao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a5e949a760f236c08ae3cebc32d8a52220b1101f5324770d40bd7d054351f6a", "submitted_at": "2026-06-12T15:58:18Z", "updated_at": "2026-06-12T15:58:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14579v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.14579", "identifiers": {"arxiv": "2606.14579"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.14579"}], "name": "VISTA: View-Consistent Self-Verified Training for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "VISTA forms GRPO comparison groups from target-preserving crops of the same GUI instance with exactly remapped target boxes. A conditional, advantage-weighted cross-view anchor supplies additional supervision without entering the group baseline, and evaluation includes view-consistency robustness.", "topics": ["grounding", "reinforcement-learning", "robustness", "data-generation"], "updated_at": "2026-06-12", "url": "https://arxiv.org/abs/2606.14579"}
{"aliases": [], "authors": ["Jingyuan Huang", "Zuming Huang", "Yucheng Shi", "Tianze Yang", "Xiaoming Zhai", "Wei Chu", "Ninghao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49b935d8993be0d03ae778375f908b68b3ec3c614a7406b6b3f6a00f2a89ef1f", "submitted_at": "2026-06-16T16:02:41Z", "updated_at": "2026-06-17T08:58:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.18101v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.18101", "identifiers": {"arxiv": "2606.18101"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18101"}], "name": "Trust the Right Teacher: Quality-Aware Self-Distillation for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes quality-aware self-distillation for GUI coordinate generation under student-produced prefixes. A correctness-aware gate suppresses coordinate-token supervision that cannot reach the target box, while teacher-probability scaling adjusts the remaining signal strength.", "topics": ["grounding", "imitation-learning", "evaluation"], "updated_at": "2026-06-17", "url": "https://arxiv.org/abs/2606.18101"}
{"aliases": [], "authors": ["Divake Kumar", "Sina Tayebati", "Devashri Naik", "Amanda Sofie Rios", "Nilesh Ahuja", "Omesh Tickoo", "Ranganath Krishnan", "Amit Ranjan Trivedi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "628d08bdbf23ae9e599bf767f0503cecedcde72dfbfcc183f9ea9f044c6f2d44", "submitted_at": "2026-06-24T12:34:28Z", "updated_at": "2026-09-29T20:55:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.25760v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.25760", "identifiers": {"arxiv": "2606.25760"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.25760"}], "name": "Uncertainty Quantification for Computer-Use Agents: A Benchmark across Vision-Language Models and GUI Grounding Datasets", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Argus benchmarks post-hoc uncertainty methods for single-step GUI grounding across open-weight and closed-source observation regimes. It tests calibration, rejection, error severity and conformal click regions, finding that uncertainty rankings transfer less reliably across model/interface regimes than across datasets for one model.", "topics": ["grounding", "evaluation", "robustness", "safety-security"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2606.25760"}
{"aliases": [], "authors": ["Chengheng Li-Chen", "Zhiqian Zhou", "Hao Chen", "Nicolas Chauvin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7a8efe4b9476d7f55ca4da17621ece653be2e33bf02be0183a339a12f5fac3ca", "submitted_at": "2026-06-24T15:35:29Z", "updated_at": "2026-06-24T15:35:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.25964v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.25964", "identifiers": {"arxiv": "2606.25964"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.25964"}], "name": "WinDOM: Self-Family Distillation for Small-Model GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "WinDOM obtains bounding-box supervision from a Windows 11 web reimplementation driven through Playwright, and studies rejection-sampling distillation for small GUI grounders. It treats the depth of the supervised cold start as a GRPO choice, comparing same-size EMA and larger same-family teachers.", "topics": ["grounding", "data-generation", "imitation-learning", "reinforcement-learning"], "updated_at": "2026-06-24", "url": "https://arxiv.org/abs/2606.25964"}
{"aliases": [], "authors": ["Sunqi Fan", "Lingshan Chen", "Runqi Yin", "Qingle Liu", "Yongming Rao", "Meng-Hao Guo", "Shi-Min Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7beb70ac91463b77a8babaf2b077c46c5072e8f86d22f4d9afc9c1e77680f309", "submitted_at": "2026-06-29T02:16:21Z", "updated_at": "2026-06-29T02:16:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29705v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.29705", "identifiers": {"arxiv": "2606.29705"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.29705"}], "name": "GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUICrafter first learns grounding from large collections of unannotated screenshots and webpages, then uses a smaller high-quality dataset for reinforcement-learning calibration. The curriculum targets reduced dependence on manually annotated GUI training data.", "topics": ["grounding", "data-generation", "reinforcement-learning", "screen-understanding"], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.29705"}
{"aliases": [], "authors": ["Chen Liu", "Ling Chen", "Hanzhang Zhou", "Liangyu Chen", "Chenglin Cai", "Xin Yu", "Steven Hoi", "Yue Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50e44e08750d4152f9b5408885f8a6a016c5f789af9712e447ee147fb25adb8d", "submitted_at": "2026-06-29T10:20:39Z", "updated_at": "2026-06-29T10:20:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30084v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.30084", "identifiers": {"arxiv": "2606.30084"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.30084"}], "name": "One Forward Beats Two: InnerZoom for Accurate and Efficient GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "InnerZoom preserves target-region cues from intermediate decoder layers and reinjects a compact evidence state during coordinate decoding. Its single-forward GUI grounder is compared with crop-and-rerun zooming and matched supervised/RL baselines for accuracy and inference cost.", "topics": ["grounding", "efficiency", "screen-understanding", "evaluation"], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.30084"}
{"aliases": [], "authors": ["Yogeswar Reddy Thota"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f3f2ae1438986c11146fec508dd61fd3e848ddcbf830f86a7dfa759280d730a", "submitted_at": "2026-06-29T03:19:40Z", "updated_at": "2026-06-29T03:19:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30697v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.30697", "identifiers": {"arxiv": "2606.30697"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.30697"}], "name": "LUMOS: A Semantic Operating-System Layer for Accessibility-Grounded AI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "LUMOS converts native accessibility metadata and browser UI structure into semantic interface blueprints with stable element identifiers and available actions. It combines live pointer-to-element grounding with an observe-act loop constrained to visible-UI primitives.", "topics": ["grounding", "screen-understanding", "tool-integration", "automation"], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.30697"}
{"aliases": [], "authors": ["Meng Chen", "Anya Ji", "Tsung-Han Wu", "Tobias Maringgele", "David M. Chan", "Alane Suhr", "Amy Pavel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0a3e641e0a71051f5fbd20f57af3db06f2b65c61025e78adcdd122f773d034b", "submitted_at": "2026-06-30T17:18:47Z", "updated_at": "2026-09-01T04:17:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.31980v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.31980", "identifiers": {"arxiv": "2606.31980"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.31980"}], "name": "DigitalCoach: Communication and Grounding Gaps in Human and Agentic Computer Use Coaching", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-30", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "DigitalCoach records expert–novice software coaching conversations aligned with screen and input events. Automated and interactive evaluations examine how model coaches differ from humans in explanation, error diagnosis and visual grounding when helping a learner operate software.", "topics": ["human-in-the-loop", "video", "evaluation", "grounding"], "updated_at": "2026-09-01", "url": "https://arxiv.org/abs/2606.31980"}
{"aliases": [], "authors": ["Chengguang Gan", "Hanjun Wei", "Yunhao Liang", "Zhixi Cai", "Qinghao Zhang", "Shiwen Ni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37cfa436d182ed8fbc8823213b9173fbfcf9fb421e8d2ae89a093a5393c374ba", "submitted_at": "2026-07-11T02:08:46Z", "updated_at": "2026-07-16T03:28:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.10079v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.10079", "identifiers": {"arxiv": "2607.10079"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.10079"}], "name": "MAG: A Web-Agent Benchmark and Harness for Multimodal Action and Guide Generation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAG jointly evaluates browser task actions and user-facing guides grounded in screenshots through element marks or pixel coordinates. Its harness covers annotation, live evaluation and joint metrics, with an expert-trajectory-augmented GRPO training study.", "topics": ["evaluation", "grounding", "reinforcement-learning", "action-prediction"], "updated_at": "2026-07-16", "url": "https://arxiv.org/abs/2607.10079"}
{"aliases": [], "authors": ["Yong Liu", "Zhenyi Zhong", "Zhanpeng Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a779b2c1ff9ba3b531d399204b0986e3affcea81c67fdd7951e3b7da343b0e17", "submitted_at": "2026-07-16T00:35:25Z", "updated_at": "2026-07-16T00:35:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.14443v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.14443", "identifiers": {"arxiv": "2607.14443"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.14443"}], "name": "Tactile: Giving Computer-Using Agents Hands and Feet", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-07-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Tactile turns accessibility, OCR and visual evidence into explicit desktop targets with executable actions and verification cues. Its execution loop chooses semantic actions or coordinate fallbacks while retaining provenance for replay and diagnosis.", "topics": ["grounding", "automation", "observability", "tool-integration", "robustness", "evaluation"], "updated_at": "2026-07-16", "url": "https://arxiv.org/abs/2607.14443"}
{"aliases": [], "authors": ["Hang Yan", "Zhangxuan GU", "Beitong Zhou", "Jiaxuan Chen", "Runze Li", "Yusong Hu", "Shuheng Shen", "Changhua Meng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfffaffc68695489e2a705515378112b232e66b6f9e8e3a9bd5df04be95bd2c9", "submitted_at": "2026-07-31T11:51:04Z", "updated_at": "2026-07-31T11:51:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.29320v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.29320", "identifiers": {"arxiv": "2607.29320"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.29320"}], "name": "MAGA: Multi-Platform Self-Fusion of GUI Agents via Structured Action Distillation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAGA distills specialized GUI agents into one policy by allocating supervision according to structured-action correctness rather than treating all response tokens equally. A training-only teacher hint supports fusion across mobile, web and desktop settings without changing student inputs.", "topics": ["imitation-learning", "grounding", "automation"], "updated_at": "2026-07-31", "url": "https://arxiv.org/abs/2607.29320"}
{"aliases": [], "authors": ["Zichuan Fu", "Shirong Wang", "Wenlin Zhang", "Guojing Li", "Yimin Deng", "Jingtong Gao", "Junjia Qi", "Hanyu Yan", "Yefeng Zheng", "Xiaopeng Li", "Wanyu Wang", "Xian Wu", "Xiangyu Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb64bb5faa37e68b35a2398ad896fc02181a4d2749a3067391e6b3bf023a0f48", "submitted_at": "2026-08-04T07:47:37Z", "updated_at": "2026-08-04T07:47:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.03270v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.03270", "identifiers": {"arxiv": "2608.03270"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.03270"}], "name": "GUI-Lens: Coarse-to-Fine Cropping for GUI Grounding with General-Purpose VLMs", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Lens uses OCR and detected interface components as coordinate references while a general-purpose VLM selects progressively enlarged screenshot crops. Instruction-conditioned checks constrain the crop and click proposals before mapping the final location back to the original screen.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2608.03270"}
{"aliases": [], "authors": ["Yu Ran", "Wentao Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "152e21cd4f1f290aafadfdcfaa10e15a03378404540a46b5f1a5464a2d432d23", "submitted_at": "2026-08-04T14:34:41Z", "updated_at": "2026-09-28T13:07:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.03740v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.03740", "identifiers": {"arxiv": "2608.03740"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.03740"}], "name": "MissClick: Execution-Aware Adversarial Attacks on Coordinate Generation in GUI Grounding Models", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "MissClick designs white-box attacks around the numerical effect of coordinate digits on executable GUI clicks. Separate untargeted displacement and targeted place-weighted objectives are evaluated across desktop, web and mobile grounding settings.", "topics": ["safety-security", "grounding", "robustness", "evaluation"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2608.03740"}
{"aliases": [], "authors": ["Yuke Li", "Xuehan Hou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f05f128cccb514987b7defb52c8f43b45d917e0e0f9570db4e75c4fe23b38533", "submitted_at": "2026-08-10T14:29:07Z", "updated_at": "2026-08-10T14:29:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09654v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.09654", "identifiers": {"arxiv": "2608.09654"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.09654"}], "name": "Hallucination-Free GUI Grounding via Regression-Free Layout-Aware Matching", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "This framework separates frozen-model instruction expansion from a trained layout-aware matcher that localizes GUI elements without coordinate regression. Binary text/icon supervision supports the matcher, with evaluations on GUI grounding and Mind2Web element-selection tasks.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2608.09654"}
{"aliases": [], "authors": ["Renshan Zhang", "Haoyang Meng", "Yixiao He", "Rui Shao", "April Hua Liu", "Liqiang Nie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87e9b3463232f606aea60df7d86626d5aa188e774b0fe515af43e1be49b1c66b", "submitted_at": "2026-08-10T15:23:52Z", "updated_at": "2026-08-10T15:23:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09723v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.09723", "identifiers": {"arxiv": "2608.09723"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.09723"}], "name": "LookAgain: Closed-Loop GUI Grounding with Visually Grounded Reflection", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "LookAgain treats a GUI coordinate as a revisable hypothesis, rendering its marker and a local screenshot patch before confirmation. Reflective supervised trajectories provide a cold start, followed by GRPO using terminal grounding correctness.", "topics": ["grounding", "self-reflection", "reinforcement-learning", "imitation-learning"], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2608.09723"}
{"aliases": [], "authors": ["Qijia Chen", "Giulio Jacucci"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cabd094270d01b20df7b029be29c5d460d10beb84e43e3e783f4f11997b74114", "submitted_at": "2026-08-22T06:30:03Z", "updated_at": "2026-08-22T06:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.21794v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.21794", "identifiers": {"arxiv": "2608.21794"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21794"}], "name": "Lexical Coupling in GUI Element Grounding: Sentence Embeddings Track Labels across Mobile and Web", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies whether text-embedding GUI grounding scores reflect semantic matching or recovery of visible labels. Same-screen ranking comparisons with lexical baselines expose differences between deployable fusion, oracle selection, and performance on label-poor targets.", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2026-08-22", "url": "https://arxiv.org/abs/2608.21794"}
{"aliases": [], "authors": ["Md Abrar Jahin", "Md Rizwan Parvez"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "970441704314fabffc46cb4a340dd6e0bde0e84ab9b7fcd2db2c9a8be9dc61d8", "submitted_at": "2026-08-22T08:04:59Z", "updated_at": "2026-08-27T07:03:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.21832v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.21832", "identifiers": {"arxiv": "2608.21832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21832"}], "name": "GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-Primitives, a paired-instruction benchmark that changes a spatial relation while holding the screenshot and anchor fixed. Its diagnostics separate failure to localize either candidate from failure to choose the candidate satisfying the relation.", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2608.21832"}
{"aliases": [], "authors": ["Jiahe Ying", "Wendong Bu", "Kaihang Pan", "Bingchen Miao", "Siyu Chen", "Wen Wang", "Xueming Jiang", "Juncheng Li", "Siliang Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "352699b04f3a9406d7cab95ce9b5fb000e3e98a28996777e75b7399a6290357a", "submitted_at": "2026-08-28T03:15:11Z", "updated_at": "2026-08-28T03:15:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27866v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.27866", "identifiers": {"arxiv": "2608.27866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.27866"}], "name": "Iron: Intent-Aligned and Retrospective Dual Learning Framework for Enhancing Generalist Virtual Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Iron, a GUI-agent training framework that aligns low-level actions with high-level intent using cycle-consistency rewards. It also repurposes failed trajectories through hindsight reconstruction and evaluates transfer to other digital tasks.", "topics": ["reinforcement-learning", "data-generation", "grounding"], "updated_at": "2026-08-28", "url": "https://arxiv.org/abs/2608.27866"}
{"aliases": [], "authors": ["Satwik Ram Kodandaram", "Monalika Padma Reddy", "Xiaojun Bi", "Jiawei Zhou", "I. V. Ramakrishnan", "Vikas Ashok"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "deb130d93bbbd6dfdc04c72d232a9dcfa1e92b0586831e5864070e7fe0de14dd", "submitted_at": "2026-09-01T00:44:19Z", "updated_at": "2026-09-01T00:44:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00524v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.00524", "identifiers": {"arxiv": "2609.00524"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.00524"}], "name": "Are We There Yet? Assessing Computer-Use Agents for Blind Users' Accessible Interaction with Desktop Applications", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "A diary study deploys the screen-reader-accessible OLLA computer-use prototype with blind users across desktop applications. Recorded commands and interaction traces support model comparisons and analysis of grounding, planning, constraint tracking and termination failures.", "topics": ["evaluation", "human-in-the-loop", "grounding"], "updated_at": "2026-09-01", "url": "https://arxiv.org/abs/2609.00524"}
{"aliases": [], "authors": ["Zhangxuan Gu", "Haoxing Chen", "Qi Qin", "Yi Xin", "Kai Gan", "Lin Liu", "Long Cui", "Xiaomei Wang", "Beitong Zhou", "Yunzhu Zhang", "Zhengwen Zeng", "Changlong Gao", "Weizhi Chen", "Rongchao Zhang", "Haoyuan Wu", "Shuheng Shen", "Changhua Meng", "Weiqiang Wang", "Jianguo Li", "Zhenzhong Lan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d7e9643c0543f088e5d2b17699c843a7643c611315f20ce92e83aea849b931a", "submitted_at": "2026-09-09T05:53:39Z", "updated_at": "2026-09-09T05:53:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13287v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.13287", "identifiers": {"arxiv": "2609.13287"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13287"}], "name": "LLaDA-UI: Bringing Block-wise Diffusion to Vision-Language GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "LLaDA-UI adapts block-wise diffusion generation to a vision-language GUI agent through multimodal pretraining and GUI supervised fine-tuning. It evaluates grounding and navigation across multiple platform settings.", "topics": ["grounding", "imitation-learning", "efficiency", "screen-understanding"], "updated_at": "2026-09-09", "url": "https://arxiv.org/abs/2609.13287"}
{"aliases": [], "authors": ["Yizhou Liu", "Fei Tang", "Yuchen Yan", "Zhengxi Lu", "Songqin Nong", "Tao Jiang", "Wenhao Xu", "Wenqi Zhang", "Weiming Lu", "Jun Xiao", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e2081a382ae815b48eb3166d18ee7c6e53afb4cd6d75538f3ff3e29f2ed5954", "submitted_at": "2026-09-14T09:56:10Z", "updated_at": "2026-09-14T09:56:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15307v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.15307", "identifiers": {"arxiv": "2609.15307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.15307"}], "name": "Learning from Reliable Negatives: Confidence-Anchored Test-Time Adaptation for GUI Grounding", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Confidence-Anchored Learning uses coordinate-token confidence to filter pseudo-supervision for GUI grounding. Its negative-only variant performs label-free test-time updates to reduce reliance on potentially incorrect positive examples.", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-09-14", "url": "https://arxiv.org/abs/2609.15307"}
{"aliases": [], "authors": ["Yao Zhao", "Aditya Shanmugham", "Swastik Roy", "Yanxun Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8dcd34beebb8a4660d8d7b9f4c76c3f86069926b9856d26b1c20d87dec83bb5", "submitted_at": "2026-09-15T04:56:17Z", "updated_at": "2026-09-15T04:56:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.16635v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.16635", "identifiers": {"arxiv": "2609.16635"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.16635"}], "name": "EchoPath: Execution-Level Replayable Memory for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "EchoPath turns validated GUI trajectories into parameterized replay procedures with explicit runtime preconditions. Image-based target matching adjusts stored coordinates, while ambiguous steps fall back to bounded repair or fresh planning.", "topics": ["memory", "workflow", "automation", "grounding", "efficiency", "robustness"], "updated_at": "2026-09-15", "url": "https://arxiv.org/abs/2609.16635"}
{"aliases": [], "authors": ["Liyang Fan", "Xinping Bi", "Yitai Li", "Shuaimin Li", "Hui Li", "Min Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccbc85f4d2598b4cb1e6e1a96b3b9184ddeba838cb610658070fd2d1ac9c8d06", "submitted_at": "2026-09-16T14:06:38Z", "updated_at": "2026-09-16T14:06:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.18690v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.18690", "identifiers": {"arxiv": "2609.18690"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.18690"}], "name": "RankGround: Efficient High-Resolution GUI Grounding via Lightweight Reranker-Guided Crop Selection", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "RankGround trains a lightweight reranker to select a promising screenshot crop before one VLM localization call. Ranking supervision uses containment and boundary-aware augmentation, with pointwise and listwise training stages.", "topics": ["grounding", "efficiency", "data-generation", "screen-understanding"], "updated_at": "2026-09-16", "url": "https://arxiv.org/abs/2609.18690"}
{"aliases": [], "authors": ["Mingfei Gao", "Rui Tian", "Haiming Gang", "Bohan Zhai", "Le Zhang", "Yuanzheng Gong", "Di Feng", "Ege Özsoy", "Kaixin Ma", "Vishwesh Kirthivasan", "Oğuzhan Fatih Kar", "Roman Bachmann", "Anders Boesen Lindbo Larsen", "Afshin Dehghan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4dfb0703133032426ff8f40ed6a2c8131025b07658c0b643603f8d50ba163d16", "submitted_at": "2026-09-18T17:59:34Z", "updated_at": "2026-09-18T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.22083v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.22083", "identifiers": {"arxiv": "2609.22083"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.22083"}], "name": "MintAct: A Unified Visual Agent for Digital Environments", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "MintAct trains a visual-agent family for grounding, navigation and tool use across mobile, desktop and web environments. Its shared infrastructure supports trajectory collection and asynchronous online RL while controlling cross-domain sampling and off-policy drift.", "topics": ["grounding", "reinforcement-learning", "data-generation", "tool-integration"], "updated_at": "2026-09-18", "url": "https://arxiv.org/abs/2609.22083"}
{"aliases": [], "authors": ["Yunxiang Li", "Xixin Wu", "Helen Meng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64936af65d0ca351b19da3f63325f2bc2a5c9e0951e5a7806cd4dd82b18784a7", "submitted_at": "2026-09-21T08:42:08Z", "updated_at": "2026-09-21T08:42:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.24277v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.24277", "identifiers": {"arxiv": "2609.24277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.24277"}], "name": "How Many Pixels Is a Digit Worth? Place-Aware Coordinate Entropy for GUI Agent Confidence Estimation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "PACE weights coordinate-digit entropy by place value to estimate confidence in GUI clicks from one forward pass. It evaluates correctness ranking and selective accuracy against repeated-sampling alternatives on ScreenSpot-Pro and ScreenSpot-v2.", "topics": ["grounding", "efficiency", "evaluation", "robustness"], "updated_at": "2026-09-21", "url": "https://arxiv.org/abs/2609.24277"}
{"aliases": [], "authors": ["Chengguang Gan", "Yunhao Liang", "QingHao Zhang", "Shiwen Ni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b2bda1c50fd47fabbe77c40a2938979a6a16b407b40f11737fc6f1fc795d090", "submitted_at": "2026-09-23T04:49:07Z", "updated_at": "2026-09-23T04:49:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27353v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.27353", "identifiers": {"arxiv": "2609.27353"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27353"}], "name": "Guides That Cause Actions: An Offline Study of Guide-Action Mutual Reinforcement in Multimodal Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebMRE pairs guidance sentences with grounded actions in an offline benchmark derived from successful WebArena traces. Controlled decoding-prefix interventions test how guide content affects action selection, while separating offline prediction metrics from live browser task success.", "topics": ["evaluation", "action-prediction", "imitation-learning", "grounding"], "updated_at": "2026-09-23", "url": "https://arxiv.org/abs/2609.27353"}
{"aliases": [], "authors": ["Zhenhao Zhang", "Zhaoyu Fan", "Haohan Ying", "Jingwen Hu", "Hancen Fan", "Junhao Zhou", "Zitian Chen", "Linchao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa2bcfa05fcbd6937af6ba3d23e892e5a105248939538dcdc9d7b015daf38593", "submitted_at": "2026-09-25T20:04:24Z", "updated_at": "2026-09-25T20:04:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.31957v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.31957", "identifiers": {"arxiv": "2609.31957"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.31957"}], "name": "CaptchaArena: A Large-Scale, Fine-Grained Dataset for Training Computer-Use Agents on Interactive CAPTCHAs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CaptchaArena, a dataset of execution-verified interactive CAPTCHA solutions with screenshot–action trajectories, reasoning annotations, and pixel masks for irregular targets. The work trains a single policy across CAPTCHA types using supervised learning followed by verifier-reward reinforcement learning.", "topics": ["data-generation", "grounding", "reinforcement-learning"], "updated_at": "2026-09-25", "url": "https://arxiv.org/abs/2609.31957"}
{"aliases": [], "authors": ["Chenyue Li", "Xiaoxiao Sun", "Yubo Deng", "Qinlin Zhao", "Serena Yeung-Levy", "Yuhui Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12976e58eb48d9c4709fb03508ac7ed7eeafffd3df675d3dccfd09bdc354f3db", "submitted_at": "2026-09-25T22:07:29Z", "updated_at": "2026-09-25T22:07:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32036v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.32036", "identifiers": {"arxiv": "2609.32036"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32036"}], "name": "ScreenHaystack: Finding Blind Zones in GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Moves controlled target icons across high-resolution interfaces to diagnose spatial blind zones in grounding models and study targeted augmentation.", "topics": ["grounding", "evaluation", "robustness"], "updated_at": "2026-09-25", "url": "https://arxiv.org/abs/2609.32036"}
{"aliases": [], "authors": ["Chuhan Zhang", "Qi Xie", "Ziyue Wang", "Jianing Yin", "Yunfan Zhou", "Dazhen Deng", "Yingcai Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f091cab38c932e08d21afa936dfe241c93c0b2139aba1d1ce7a6d1513ad5da1a", "submitted_at": "2026-09-26T08:58:50Z", "updated_at": "2026-09-26T08:58:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32385v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.32385", "identifiers": {"arxiv": "2609.32385"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32385"}], "name": "DashAct: A Progressive Diagnostic Benchmark for GUI Agents in Interactive Dashboard Analysis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Diagnoses dashboard-agent failures by progressively restoring interaction context, next-action support, and visual target information within the same task.", "topics": ["evaluation", "grounding", "planning"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32385"}
{"aliases": [], "authors": ["Keliang Li", "Heng Wang", "Chen Hu", "Daxin Jiang", "Hong Chang", "Shiguang Shan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88daacf5fda62f59a08b504a633ad984ee1d03791bba24e89c223c0686adf34e", "submitted_at": "2026-09-27T15:11:30Z", "updated_at": "2026-09-27T15:11:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33646v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.33646", "identifiers": {"arxiv": "2609.33646"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.33646"}], "name": "Probe to Act: Elevating Browser-Use Agent via Active Visual Probing", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Probe to Act, which lets browser agents request visual evidence linking DOM handles and screen regions before committing an operation. It retains selected interaction evidence as memory and studies both prompted use and distillation into open models.", "topics": ["grounding", "memory", "web-navigation"], "updated_at": "2026-09-27", "url": "https://arxiv.org/abs/2609.33646"}
{"aliases": [], "authors": ["Ziqi Zhang", "Shaohui Li", "Bing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7efd2eafa1984d7c8f4bf6ce79a94cd2ca4189da330bdd056b4be9c07f7f9dc1", "submitted_at": "2026-09-28T04:02:07Z", "updated_at": "2026-09-28T04:02:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35904v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.35904", "identifiers": {"arxiv": "2609.35904"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35904"}], "name": "Structured Interaction, Visual Localization, and Robust Execution for Complex Web Tasks: A Technical Report on the WebRetriever Challenge", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a WebRetriever agent that combines semantic webpage operations with visual localization when structured representations are insufficient. Hierarchical context management and fault-aware execution support processing tasks across concurrent browsers.", "topics": ["web-navigation", "grounding", "robustness", "memory"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.35904"}
{"aliases": [], "authors": ["Shengtian Yang", "Ziyu Xiong", "Kaibing Yang", "Guangfeng Cai", "Yewen Li", "Peng Jiang", "Gai Kun", "Qingpeng Cai", "Lei Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d45cee8117e302aa190d38c8298bf479936a5a27aa9b85716a76a2a90499069c", "submitted_at": "2026-09-29T07:51:57Z", "updated_at": "2026-09-29T07:51:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36939v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.36939", "identifiers": {"arxiv": "2609.36939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36939"}], "name": "SCA: Spatial Credit Assignment for Reinforcement Learning of GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Spatial Credit Assignment, which uses click coordinates to refine group-relative reinforcement-learning updates for GUI grounding. It distinguishes failed clicks by spatial evidence while leaving the deployed action policy unchanged.", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.36939"}
{"aliases": [], "authors": ["Cheng Yang", "Yifan Wu", "Yutao Huang", "Zhaohua Zhang", "Beiduo Chen", "Muxi Chen", "Chenchen Zhao", "Hexuan Deng", "Haolin Yang", "Geyuan Zhu", "Sa Zhu", "Jianhuan Zhuo", "Qiuyong Xiao", "Jianhao Ruan", "Yiran Peng", "Jiayi Zhang", "Tian Ye", "Xinlei Yu", "Tianwen Jiang", "Jihong Zhang", "Yuyu Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecd40b78c6de41abb442705b993aad8ebf7b35741536a11e00ebbd60140d1f18", "submitted_at": "2026-10-01T07:22:53Z", "updated_at": "2026-10-01T07:22:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01215v1"}}, "environments": ["desktop", "linux", "windows", "macos", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.01215", "identifiers": {"arxiv": "2610.01215"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.01215"}], "name": "AutoGUIWorld: Image Generators as Visual World Models for GUI Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AutoGUIWorld, which combines task planning and image editing to synthesize action-grounded GUI trajectories without running the depicted software. It filters generated transitions and tests whether training on them improves agents operating real desktop and scientific applications.", "topics": ["world-models", "data-generation", "grounding"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2610.01215"}
{"aliases": [], "authors": ["A. Said Gurbuz", "Ahmed Nassar", "Sunghwan Hong", "Marc Pollefeys", "Peter W. J. Staar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd66468c2670ec22216e1592719a05e0b56af69606237a50f2067a340172553e", "submitted_at": "2026-10-01T18:00:05Z", "updated_at": "2026-10-01T18:00:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02320v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.02320", "identifiers": {"arxiv": "2610.02320"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02320"}], "name": "DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Generates dense desktop supervision by varying real application scenes and combining screenshots, accessibility trees, and window geometry.", "topics": ["data-generation", "grounding"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2610.02320"}
{"aliases": [], "authors": ["Kaiyu Wu", "Beichen Zheng", "Weiyao Huang", "Keze Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d20eb090847d25deb8f69181caa6f964661bbf167223b5572cf3e1cafa52bbd", "submitted_at": "2026-10-04T12:49:01Z", "updated_at": "2026-10-04T12:49:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.05185v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.05185", "identifiers": {"arxiv": "2610.05185"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.05185"}], "name": "Recurrent Latent Visual Search for GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ReLaViS, which performs recurrent visual search inside a grounding model by repeatedly querying screenshot tokens and feeding the resulting spatial evidence back into the model. Coarse-to-fine supervision is constructed from existing interface-element annotations.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2026-10-04", "url": "https://arxiv.org/abs/2610.05185"}
{"aliases": [], "authors": ["Aadi Chauhan", "Arthur Ilyasov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81342daf25b0fb26f6481fcf836c3d958e26c2358be96eadb51e6914ea87e43c", "submitted_at": "2026-10-05T21:50:50Z", "updated_at": "2026-10-05T21:50:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07444v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.07444", "identifiers": {"arxiv": "2610.07444"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.07444"}], "name": "Decoupling What from Where: How Should a Small GUI Grounding Model Receive the Action Type?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares ways to condition a small GUI grounding model on action type under matched training conditions on Android in the Wild. The study traces apparent gains to a coordinate-serialization choice and reports limited evidence that the gains persist with predicted action types.", "topics": ["grounding", "evaluation", "robustness"], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2610.07444"}
{"aliases": [], "authors": ["Wanjing Han", "Levi Taiji Li", "Mu Zhang", "Yue Jiang", "Guanhong Tao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b6ef7e785ad509ebd13fdb3da97fc583ee4c28dd42b4eab842c6cd88d7ef51f", "submitted_at": "2026-10-07T00:05:53Z", "updated_at": "2026-10-07T00:05:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.09240v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.09240", "identifiers": {"arxiv": "2610.09240"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.09240"}], "name": "Adversarial Images Hijack Web Agents from Visual Grounding to Browser Execution", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebMirage, a visual red-teaming framework that models webpage element competition and action post-processing to test whether manipulated grounding decisions lead to executed browser actions.", "topics": ["safety-security", "grounding", "evaluation"], "updated_at": "2026-10-07", "url": "https://arxiv.org/abs/2610.09240"}
{"aliases": [], "authors": ["Shangquan Sun", "Ruoyu Chen", "Zhirui Liu", "Xiang Liu", "Chenghao Qian", "Wenqi Ren", "Xiaochun Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f7effcf5e101cfcc7e229fe4aab40d07401b1dd98cf5ce3b566fef863a42190", "submitted_at": "2026-10-09T01:50:26Z", "updated_at": "2026-10-09T01:50:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12866v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.12866", "identifiers": {"arxiv": "2610.12866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12866"}], "name": "CaptchaAgent: Verifiable Visual Reasoning for CAPTCHA Solving", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents CaptchaAgent, which builds an explicit visual representation of objects and relations, refines ambiguous regions, and checks counterfactual consistency before acting on interactive CAPTCHAs.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.12866"}
{"aliases": [], "authors": ["Ruotian Zhang", "Yuhui Zhang", "Junzhe Huang", "Yu Xin", "Hehe Fan", "Xiaoxiao Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "010e8ebc657dccad670ea6d17985e9c3c1680017f2516eaa86b7ea9db5660990", "submitted_at": "2026-10-09T09:20:02Z", "updated_at": "2026-10-09T09:20:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13268v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.13268", "identifiers": {"arxiv": "2610.13268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13268"}], "name": "ConClick: Element-Wise Contrastive Learning Improves GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes recurring distractor clicks in GUI grounding and introduces ConClick, a contrastive training method built from target–distractor pairs. Its ReClick inference procedure revisits candidate regions and compares competing predictions.", "topics": ["grounding", "data-generation", "robustness"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.13268"}
{"aliases": [], "authors": ["Jianlin Chen", "Wenhui Chen", "Chi Man Vong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6c60bb4d3b7bb2e2cba92201924d8da4f5e18ce5553c81fc9d5c6fb62b84751", "submitted_at": "2026-10-11T16:18:11Z", "updated_at": "2026-10-11T16:18:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15738v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.15738v1"], "reason": "Publisher document declaration only; no code, data or weights license inferred."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:10Z", "role": "primary", "sha256": "6d8c0463b6d50320246021f0394d0e39c9fd839f1ef1f742c8cdfa231872b059", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:31:34.620763Z", "role": "primary", "sha256": "7ec947ceb462b5fd23a4a67df64e931e427224ef721b681d3fc92c1d4379449f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.15738v1"}], "id": "arxiv:2610.15738", "identifiers": {"arxiv": "2610.15738"}, "kind": "paper", "last_verified_at": "2026-10-20T15:31:34.620763Z", "licenses": {"code": null, "data": null, "document": {"basis": "The retained version-specific arXiv HTML infobox explicitly declares License: CC BY 4.0 for this document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:31:34.620763Z", "url": "https://arxiv.org/html/2610.15738v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.15738"}], "name": "Visible but Not Avoided: State-Dependent Re-Proposal of Rejected Candidates in Catalog-Based LLM Optimization Loops", "notes": "The downstream test uses one screenshot-only agent, nine synthetic apps and 60 cached development items, with TAP-tag or text-answer outputs; this is not demonstrated live-app completion. Two catalogs reword the same 40 edits. Re-proposal effects do not establish internal causal mechanisms or task-success gains. Main serving providers were unpinned; rationale labels came from LLM annotators. No independent reproduction.", "published_at": "2026-10-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies repeated rejected edits while optimizing a screenshot-only GUI agent, contrasting static probes with reached loop states and testing how rejection wording changes subsequent edit proposals.", "topics": ["evaluation", "grounding", "training-free"], "updated_at": "2026-10-11", "url": "https://arxiv.org/abs/2610.15738"}
{"aliases": [], "authors": ["Jingchao Wang", "Dingjiang Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa0e0ab6eb19ea55e65ccb354ae515f6552cd03dd688c6f43cdc6e2d59b11824", "submitted_at": "2026-10-12T06:09:02Z", "updated_at": "2026-10-12T06:09:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16234v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.16234", "identifiers": {"arxiv": "2610.16234"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16234"}], "name": "Seg2Click: Learning Pixel-Level Action Masks for GUI Visual Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Recasts GUI grounding as predicting a mask of valid click locations from bounding-box supervision, then derives a click point with a deterministic selection rule.", "topics": ["grounding", "screen-understanding"], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.16234"}
{"aliases": [], "authors": ["Jiahao Lu", "Yize Shen", "Chuanyuan Tan", "Mianqiu Huang", "Shicheng Fang", "Xinchi Chen", "Yining Zheng", "Yuxin Wang", "Xipeng Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73073d2d533307cd71b05afef2441a8a7b8f198f804f11f83536d4297db2491a", "submitted_at": "2026-10-12T12:52:21Z", "updated_at": "2026-10-12T12:52:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.16720", "identifiers": {"arxiv": "2610.16720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16720"}], "name": "Learn from Move: the Next Step for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces LatentGUIWorld to test learning interface semantics and dynamics through interaction, and studies finer movement actions with intermediate visual feedback and exploration demonstrations.", "topics": ["evaluation", "grounding", "reinforcement-learning"], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.16720"}
{"aliases": [], "authors": ["Minghua Chen", "Xingming Liao", "Weiqiang Zhong", "Zhuowei Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b1786bba4f48ca54166d49e3d98d55aaef52eb6e2233f282d8d94d54dfb5ed2c", "submitted_at": "2026-10-13T06:20:38Z", "updated_at": "2026-10-13T06:20:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17792v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.17792", "identifiers": {"arxiv": "2610.17792"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17792"}], "name": "MHVF: Training-Free GUI Grounding via Verification-Guided Multi-Hypothesis Search", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MHVF, a training-free grounding method that explores several candidate screen regions, verifies them in local views, and uses those checks to guide further search.", "topics": ["grounding", "screen-understanding", "robustness"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.17792"}
{"aliases": [], "authors": ["Qingsong Xie", "Jiahao Tian", "Jiacheng Tang", "Zhifeng Li", "Madalina Ciobanu", "Navan Preet Singh", "Qingqing Mao", "Ritankar Das"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08d90bcec690b238be96207602ca694759c8672fecbed85e30372dd314e4006e", "submitted_at": "2026-10-13T19:37:49Z", "updated_at": "2026-10-13T19:37:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19100v1"}}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.19100", "identifiers": {"arxiv": "2610.19100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19100"}], "name": "GroundGUI: GUI Grounding by High-Quality Data and Distillation-to-Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ScreenRef screenshot grounding data and a training sequence from supervised learning through teacher-guided distillation to reinforcement learning for GroundGUI-8B.", "topics": ["grounding", "data-generation", "reinforcement-learning"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.19100"}
{"aliases": [], "authors": ["Yuke Wang", "Benlei Cui", "Shen Pang", "Xuemei Dong", "Longtao Huang", "Hui Xue", "Yuwen Zhai", "Junjie Li", "Jingqun Tang", "Haiwen Hong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da334d943d98b4bf9915c6ca28aab734d736f376e67cbab950aa1af65f540ebf", "submitted_at": "2026-10-14T01:30:03Z", "updated_at": "2026-10-14T01:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19384v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.19384v1"], "reason": "Paper-specific document declaration only; no code, data or weights permissions are inferred."}}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:02:16.329923+00:00", "role": "primary", "sha256": "ca995ec330f62a43d3d4fbbf28dfe83c5fa8b05f5c1e446929493409760268d1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.19384v1"}], "id": "arxiv:2610.19384", "identifiers": {"arxiv": "2610.19384"}, "kind": "paper", "last_verified_at": "2026-10-20T19:02:16.329923+00:00", "licenses": {"code": null, "data": null, "document": {"basis": "The retained exact-v1 arXiv HTML infobox explicitly declares License: CC BY 4.0 for this document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:02:16.329923+00:00", "url": "https://arxiv.org/html/2610.19384v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19384"}], "name": "MMA-SafetyBench: A Benchmark for Multimodal Agent Safety Evaluation", "notes": "ASR measures predicted coordinates inside an annotated target or compromised generated content; actual clicks, environment changes and downstream harm are not verified. SCR additionally counts adopted adversarial intent or task/schema disruption. Clean-task and attacked-trajectory denominators differ, including 42 clean desktop tasks versus 209 adversarial desktop trajectories. Payloads and injection points are fixed, and neutral prompts change both compliance cues and instruction-priority guidance. GPT-5 judging has a 300-trajectory human-label check, with borderline disagreements. Android-style mobile dialogs do not establish an Android runtime. Code/data links were identified but not inspected; no independent reproduction.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds trajectory-guided visual-injection tests for web, desktop and mobile observations, separating target-matching action predictions from broader semantic compromise.", "topics": ["evaluation", "safety-security", "grounding"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.19384"}
{"aliases": [], "authors": ["Chengguang Gan", "Yunhao Liang", "Hanjun Wei", "Qinghao Zhang", "Zhixi Cai", "Shiwen Ni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47488d3b5450753b4e1a4b0991e50dc92fbab197a9b54bd8d4e3e1a3d762ee96", "submitted_at": "2026-10-14T12:39:51Z", "updated_at": "2026-10-14T12:39:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20107v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.20107", "identifiers": {"arxiv": "2610.20107"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20107"}], "name": "Why Generate a Click? Plug-in Decision Heads for Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adds decision heads to a frozen vision-language generator for selecting web actions and targets, with a confidence gate that falls back to text generation.", "topics": ["efficiency", "grounding", "web-navigation"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.20107"}
{"aliases": [], "authors": ["Shuo Wang", "Sifan Zhou", "Xiaobo Wang", "Tianyu Fu", "Chenxu Zhao", "Anyang Su", "Hanning Wang", "Ping Jiang", "Minghui Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f9d3ce7dffa8c726ce8129083c8d6c580449d1a32d87c7cd3aaade4fe6cf59b", "submitted_at": "2026-10-14T15:53:48Z", "updated_at": "2026-10-14T15:53:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20390v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.20390", "identifiers": {"arxiv": "2610.20390"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20390"}], "name": "STP-GUI: Spatially-Aware Token Pruning for Efficient Web-GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces STP-GUI, which selects spatially coherent screenshot windows conditioned on the instruction to reduce visual tokens while retaining useful interface structure.", "topics": ["efficiency", "grounding", "screen-understanding"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.20390"}
{"aliases": [], "authors": ["Fuyao Zhang", "Wenwen He", "Jiaming Zhang", "Yurong Hao", "Xuan Wang", "Guowei Guan", "Che Wang", "Longtao Huang", "Wei Yang Bryan Lim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "41178d19a23653c0f49423b8f725584bded7ab8819a1a46e83a062794460b4cb", "submitted_at": "2026-10-19T11:30:48Z", "updated_at": "2026-10-19T11:30:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26203v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:23.872790+00:00", "role": "primary", "sha256": "28848f8f597b2a8865ad98f5aa1e4401e15a49f92b33f7bce4d4f044c168ada0", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26203v1"}], "id": "arxiv:2610.26203", "identifiers": {"arxiv": "2610.26203"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26203"}], "name": "Focused Omni-Modal GUI Agents", "notes": "Reviewed the complete retained abstract and selected v1 method/evaluation sections. OmniGUI uses episode-disjoint reference-action evaluation; Type Match, Exact Match and Goal Progress are not presented as independently demonstrated live task-success gains. The Android role is explicit in Appendix D.1. No reproduction or artifact-release verification.", "published_at": "2026-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "FocusGUI learns step-specific weights over screenshot, audio, video and interaction-context channels, injecting a compact residual into a frozen decoder without pruning its inputs.", "topics": ["screen-understanding", "grounding", "imitation-learning", "interaction-history", "mobile-navigation"], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26203"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "1973f702c08ca9637ce03bc28fc34188be05826e9d43f9992b08e804f474b1a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/apple-aiml-research/ml-ferret/main/ferretui/README.md"}], "id": "github:apple/ml-ferret", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/apple/ml-ferret"}], "name": "Ferret-UI", "notes": "The repository has moved from apple/ml-ferret to apple-aiml-research/ml-ferret; evidence uses the dedicated ferretui component.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/apple-aiml-research/ml-ferret/main/ferretui/README.md", "target_id": "arxiv:2404.05719", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screen understanding research within the Ferret model family.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://github.com/apple/ml-ferret"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}], "id": "github:ariaui/aria-ui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AriaUI/Aria-UI"}], "name": "Aria-UI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mixture-of-experts GUI grounding model with 3.9B activated parameters that can use text or interleaved image history, with released training data and an AndroidWorld agent.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/AriaUI/Aria-UI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "0c4feb17b657c0c9a71d7e0c5e2142abbd19bb5cff22e6cd3b6b082644c98ee7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/LICENSE"}], "id": "github:bytedance/ui-tars", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytedance/UI-TARS"}], "name": "UI-TARS", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/README.md", "target_id": "arxiv:2501.12326", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Native GUI agent models; UI-TARS-1.5 adds reinforcement-learned reasoning before acting and an open 7B checkpoint, and UI-TARS-2 extends to games, code and tool use.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://github.com/bytedance/UI-TARS"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["READ text answer", "ACT selector, text/label, accessible role/name, DOM index, Stagehand reference or screenshot point, according to the representation"], "counts": [], "evaluation": "READ uses normalized exact match against an answer or allowed alias. ACT resolves predicted identifiers to ground-truth elements, or checks screenshot points inside the target box. Representation-token accounting subtracts fixed-prompt tokens; screenshots use billed input usage minus the fixed prompt. These are answer/target correctness and token metrics, not completed-task success.", "evidence_urls": ["https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/README.md", "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/METHODOLOGY.md", "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/src/wpbench/scoring.py"], "focus": ["Browser-representation token use", "Factual READ accuracy", "ACT target-element selection"], "notes": "Fixed local-page observations; the reported run uses claude-sonnet-5 and a shared prompt, without scrolling or an agent loop. Browser Use disables its viewport cutoff and sends untruncated text; Stagehand uses a separate matched browser load. Short-string representability labels are approximate. Offline does not mean API-free. Third-party snapshot redistribution remains unresolved; results were not reproduced.", "observations": ["Top-document HTML", "Markdown converted from HTML", "Playwright accessibility snapshot", "Browser Use indexed DOM", "Stagehand hybrid snapshot", "Viewport screenshot"], "setting": "offline"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"], "reason": "The inspected root LICENSE contains complete MIT terms for repository software. The methodology explicitly leaves redistribution permission for captured third-party practice-site pages unresolved; code MIT is not a blanket snapshot/data license."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:30.874397Z", "role": "primary", "sha256": "14775f6e3daba69d4684380fd33bf717bdc3093b65be796eb863bbd27d0c0bab", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "environments", "details.benchmark_protocol.focus", "details.benchmark_protocol.setting", "details.benchmark_protocol.observations", "details.benchmark_protocol.actions", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:05:35.063004+00:00", "role": "primary", "sha256": "3e22e1cbfa47bcf98f6f70a5f6f162a31a509a9f2091a84679bb5231977e29d9", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Aweb-agents+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:50:54.545491Z", "role": "primary", "sha256": "fd30e94e78e7167a49579922259fde1c7d7f8259e9f00afba15faf3a71966a94", "supports": ["id", "url", "notes"], "url": "https://github.com/charlypoly/web-context-bench-python"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:30.131541Z", "role": "primary", "sha256": "1c3c0cbf9f9d9f7cd9e7a2f027ecb217fd5c16fa157916281a15bb7f403f6f1c", "supports": ["summary", "environments", "topics", "scope", "notes", "licenses.data", "details.benchmark_protocol.focus", "details.benchmark_protocol.setting", "details.benchmark_protocol.observations", "details.benchmark_protocol.actions", "details.benchmark_protocol.evaluation", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/METHODOLOGY.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:28.567861Z", "role": "primary", "sha256": "8cada5d71b5a61e7aff2e1ac2b951c266eb2db7c082c06c5a13aa127d42b3fc5", "supports": ["summary", "topics", "scope", "notes", "details.benchmark_protocol.evaluation", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/src/wpbench/scoring.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:41.499735Z", "role": "primary", "sha256": "9b64ef672f5091b6074d02502f6d3c161d94308a3d9d142d7437dcddf611bbc4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"}], "id": "github:charlypoly/web-context-bench-python", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-20T11:51:30.874397Z", "licenses": {"code": {"basis": "Complete MIT license text in the repository root at the inspected commit; repository code only.", "declared": "MIT", "retrieved_at": "2026-10-20T11:51:41.499735Z", "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/charlypoly/web-context-bench-python"}], "name": "Web Context Bench (Python)", "notes": "This is a single-snapshot representation evaluation on local fixtures and captured pages, not a live multi-step agent benchmark. Its shared-prompt, one-model setup differs from complete Browser Use or Stagehand agents. The methodology notes unresolved redistribution licensing for third-party practice-site snapshots; MIT is recorded for repository code only. Selected scoring code was read, but no models, browser tasks or claimed results were run or reproduced.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Local-page benchmark comparing six browser representations on token use, factual reading and target-element selection, with deterministic answer scoring.", "topics": ["evaluation", "grounding", "efficiency", "screen-understanding"], "updated_at": null, "url": "https://github.com/charlypoly/web-context-bench-python"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"], "reason": "The inspected root LICENSE is complete MIT text for repository code. The local data README says same as parent repository, but the linked Hub dataset/model cards and artifacts were not inspected; their licenses are not inferred."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:34.292633Z", "role": "primary", "sha256": "7cda5207d8882bc8dedaefb93d2ccf54b310354917283b8f4890fffdc5764958", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "environments", "relations"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:04:53.470789+00:00", "role": "primary", "sha256": "3d76d934e98641746085772b98ee17059749f980a616435498a7ad483bd4cec1", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-grounding+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:50:58.795155Z", "role": "primary", "sha256": "08b89c5e6143afa4c0cdde95f8ec016d9033ee004f7631fadd4457144c2b87c9", "supports": ["id", "url", "notes"], "url": "https://github.com/ChenghengLi/WinDOM"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:41.278660Z", "role": "primary", "sha256": "09488140d35408cc11bb9c12b1bb435b21814cf6525937a8309475c9326c4004", "supports": ["environments", "notes", "licenses.data"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/data/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:45.091369Z", "role": "primary", "sha256": "0ee33b1e416ea4c5d8834688e6f42fbd15ef493f1b82398d3bb985838cbdffb9", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/src/windom/training/sfd_4b.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:48.511963Z", "role": "primary", "sha256": "c481b2f91f2d82bf801632da6ab6cce01bc42c780104d743156c272942d2cef3", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/src/windom/evaluation/screenspot.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:44.940106Z", "role": "primary", "sha256": "d5a37ff86df9b231440ab1bcf3ecf77304a0a53e772182cdb622a3ab5d043447", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"}], "id": "github:chenghengli/windom", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T11:51:34.292633Z", "licenses": {"code": {"basis": "Complete MIT license text in the repository root at the inspected commit; repository code only.", "declared": "MIT", "retrieved_at": "2026-10-20T11:51:44.940106Z", "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ChenghengLi/WinDOM"}], "name": "WinDOM code", "notes": "The Windows-style environment is win11React in Playwright, not native Windows control. The repository documents its predecessor trajectory pipeline but explicitly omits the final corpus harvester. The inspected ScreenSpot evaluator defaults to a privileged ground-truth-center demonstration; --no-demo and --official-prompt select other modes, so evaluation results require their prompt and coordinate settings. Linked datasets, weights and reported scores were not independently verified; MIT applies to inspected repository code.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/README.md", "target_id": "arxiv:2606.25964", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI-grounding training and evaluation code with same-family distillation, GRPO recipes and point-in-box scoring, using a Windows-style web environment.", "topics": ["grounding", "imitation-learning", "reinforcement-learning", "evaluation"], "updated_at": null, "url": "https://github.com/ChenghengLi/WinDOM"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"], "reason": "MIT declaration in the author repository LICENSE; separate data, weights and document scopes are not inferred."}}, "repository_id": 1079441024, "repository_revision": "2d69eeaf0ac44210c76dcab2d08b4d4610be236c"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:53.536297+00:00", "role": "primary", "sha256": "cbb08803922b4b7736979ca0a5fe552b111fcd428fa9317401bc090b14bb4e54", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:30.768857+00:00", "role": "primary", "sha256": "ff80afd35c79d8dd27dd7b9284ce082fe238c13a113aa19400ca0dc2d18ff194", "supports": ["identity", "details.repository_id"], "url": "https://api.github.com/repos/ghadinehme/VideoCAD"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:52.564274+00:00", "role": "primary", "sha256": "9ded7c6ef25b88879e9d46939ca7ff3bef97362e76dbdb808e57c67008541557", "supports": ["release_status", "details.repository_revision", "notes"], "url": "https://api.github.com/repos/ghadinehme/VideoCAD/git/trees/2d69eeaf0ac44210c76dcab2d08b4d4610be236c?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:34.311949+00:00", "role": "primary", "sha256": "a303d95a811331cd6ee23b0340f1b9ef2dd996835bfbc66656a871ff94eead50", "supports": ["environments", "summary", "relations", "notes"], "url": "https://arxiv.org/html/2505.24838v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:24.765415+00:00", "role": "primary", "sha256": "8e02753a1016dc67c97b395934313027b5f91b74209c7a24438de70fdc19cad1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"}], "id": "github:ghadinehme/videocad", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T12:02:53.536297+00:00", "licenses": {"code": {"basis": "The author repository at the inspected commit contains the standard MIT license. This declaration does not establish terms for externally hosted data or weights.", "declared": "MIT", "retrieved_at": "2026-10-18T12:03:24.765415+00:00", "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ghadinehme/VideoCAD"}], "name": "VideoCAD training and evaluation code", "notes": "The paper links this author repository. Training, preprocessing, model and evaluation source files are released; runtime behavior has not been independently reproduced. Onshape is a browser-based CAD environment; supported operating systems are not inferred. The full dataset and a checkpoint are linked separately, and their payloads were not downloaded for this review. MIT is recorded only for this repository code; external data and weight licenses remain unknown.", "published_at": null, "relations": [{"evidence_url": "https://arxiv.org/html/2505.24838v2", "target_id": "arxiv:2505.24838", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "VideoCADFormer training, action-log preprocessing and evaluation code for learning CAD interface action sequences from screen recordings of browser-based Onshape.", "topics": ["grounding", "imitation-learning", "evaluation"], "updated_at": null, "url": "https://github.com/ghadinehme/VideoCAD"}
{"aliases": [], "authors": [], "details": {"availability": "Public archived TFRecord annotation files for both downstream tasks.", "dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"], "formats": ["TFRecord"], "modalities": ["text", "bounding boxes", "UI hierarchy metadata", "Rico image identifiers"], "notes": "Original UI images are retrieved separately by Rico IDs. No release-specific sample count is established.", "splits": ["AppSim train/dev/test", "RefExp train/dev/test"], "supervision": ["similar UI element retrieval", "referring-expression component retrieval"]}, "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"], "reason": "The README explicitly says the datasets are licensed under CC BY 4.0; original Rico materials keep their source terms."}}, "repository_revision": "eceafbde7567959899bfb7f8d9e6fd71a52aaa12"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "45d1ca0f5c727cf58b2671dc5579f0628f4d78e07b977a5d8b018295c6fe1f58", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations", "details.dataset_profile", "licenses.data"], "url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "24a7838d425bdae363a7d5872702996b22068c1f0e47be274fdfbc4b0e20144f", "supports": ["identity"], "url": "https://api.github.com/repos/google-research-datasets/uibert"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "9a6fc19c6b407309943fe645f4addb3d032f153856976c4f8d7a8cfb6d2ffe6b", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/google-research-datasets/uibert/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "f752ebfb2598b7f848d343f428502b28e0f591f27849ab707f14bee398d4eb8d", "supports": ["environments", "notes"], "url": "https://interactionmining.org/rico"}], "id": "github:google-research-datasets/uibert", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": {"basis": "The README explicitly says the datasets are licensed under CC BY 4.0; original Rico materials keep their source terms.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-17T12:30:54Z", "url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/uibert"}], "name": "UIBert AppSim and RefExp annotations", "notes": "The archived repository retains TFRecord train/dev/test files for AppSim and RefExp. These are annotation and UI-structure records with Rico image identifiers, not a republished complete screenshot collection or a UIBert checkpoint. The README’s 72k figure describes the upstream Rico source, not the number of released AppSim/RefExp examples; no sample total is inferred. The publisher explicitly assigns CC-BY-4.0 to these datasets.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md", "target_id": "url:0b134b082a7bd507fc6a", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md", "target_id": "arxiv:2107.13731", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Two Rico-derived UI component retrieval datasets: AppSim pairs semantically similar elements across screens, while RefExp links referring expressions to target elements.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": null, "url": "https://github.com/google-research-datasets/uibert"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "68cb0deb8a2eb49fb9236ae3d885b043d240dded0fce31cf7425130a56ff5819", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/HeimingX/TAG/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "a33bee17685d64d23fb4d65b30b9c52f3a6dd5b21789d7b3509eb2e15145de6f", "supports": ["identity"], "url": "https://api.github.com/repos/HeimingX/TAG"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "fac4f5e38011bbcfef916063d495c21d49830759af4e8713a60d7bae992d2710", "supports": ["release_status"], "url": "https://api.github.com/repos/HeimingX/TAG/git/trees/main?recursive=1"}], "id": "github:heimingx/tag", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/HeimingX/TAG"}], "name": "TAG attention-driven GUI grounding", "notes": "The release contains grounding and benchmark evaluation code. The inspected root repository metadata does not declare a code license. It uses external pretrained models rather than publishing a new trained checkpoint, and the README restricts redistribution of unzipped benchmark data.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/HeimingX/TAG/HEAD/README.md", "target_id": "arxiv:2412.10840", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Grounding and evaluation code that aggregates a pretrained vision-language model’s attention maps to localize requested interface elements without fine-tuning.", "topics": ["grounding", "evaluation", "training-free"], "updated_at": null, "url": "https://github.com/HeimingX/TAG"}
{"aliases": [], "authors": [], "details": {"availability": "Public benchmark-adapter source and running instructions.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"], "reason": "Inspected repository LICENSE; not a license for linked benchmark datasets, pretrained weights or services."}}, "repository_revision": "9ceda5070d5664dd9376d5f8e7b43f02188e01ff"}, "environments": ["web", "android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:13Z", "role": "primary", "sha256": "edcaa6f8355be6a8774ebeeb9bdb67648e8eeddb813d6be0fa2433116d9def36", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:12Z", "role": "primary", "sha256": "e372458e48d0ebbbe0f7887b66e2d5bb4db612fd96db4dc53404c0319891f7e8", "supports": ["identity"], "url": "https://api.github.com/repos/huggingface/screensuite"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:05Z", "role": "primary", "sha256": "93b685a13bfba4d9569dbfc0418de93c7b9f8a3f7c101af053aadeebbdf8b2c0", "supports": ["release_status"], "url": "https://api.github.com/repos/huggingface/screensuite/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:06Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"}], "id": "github:huggingface/screensuite", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:12:13Z", "licenses": {"code": {"basis": "Inspected repository LICENSE; not a license for linked benchmark datasets, pretrained weights or services.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T09:14:06Z", "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/huggingface/screensuite"}], "name": "ScreenSuite", "notes": "The publisher states that the suite compares the underlying multimodal models rather than full agent implementations. Public source includes benchmark adapters and registry code. Multi-step environments require Docker and KVM; constituent datasets, remote model services and OSWorld account-dependent tasks retain their own setup and terms. Apache-2.0 is scoped to this repository code.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A shared evaluation harness for GUI-model perception, offline action prediction and multi-step interaction, using simple smolagents-based agents across existing benchmarks.", "topics": ["evaluation", "grounding", "action-prediction"], "updated_at": null, "url": "https://github.com/huggingface/screensuite"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5703c892aaf99d250bf59edf52d359560a80d443f2be13c7aa76af0c1a7f1d67", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md"}], "id": "github:inclusionai/ui-venus", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/inclusionAI/UI-Venus"}], "name": "UI-Venus", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2508.10833", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2602.09082", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2609.00028", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI agent model family; UI-Venus-2, at 9B and 27B, acts on mobile, web and desktop in one end-to-end agent, trained with keypoint-based multi-judge task verification.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://github.com/inclusionAI/UI-Venus"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "4b68c82a2dd0063a450bde34c8bffbd931565a91a2982a5b47be7e711815b284", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/LICENSE"}], "id": "github:infixai/infigui-g1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/InfiXAI/InfiGUI-G1"}], "name": "InfiGUI-G1", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md", "target_id": "arxiv:2508.05731", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "InfiGUI-G1 3B and 7B grounding models trained with Adaptive Exploration Policy Optimization, which samples several candidate answers per pass under an adaptive exploration reward.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/InfiXAI/InfiGUI-G1"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "2316acbef1207925b10b4770e14bdf69f6f4b6b3bc0ddb12e6be107d6f1425da", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "57bcc2eb985fa6036bf2bde7f28c4fd359c6b3479a658fea768162cac2a75ba5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/LICENSE"}], "id": "github:jcwang0602/seg2click", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:46:05Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jcwang0602/Seg2Click"}], "name": "Seg2Click", "notes": "The repository provides training, evaluation, and a custom mask decoder. Its code license does not determine the licenses of the five model checkpoints. No model execution was performed.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training and inference implementation that predicts an executable-region mask and derives a GUI click from that mask.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://github.com/jcwang0602/Seg2Click"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "eb2a2767741b53e5a6659ef4b0cc325ac18ffdc4222deef1d0e7671089fcaa39", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "a145c4ec8fd43f59ebe97708b6ee4cef8112959036957eaa4739b8da9a297e8a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/LICENSE"}], "id": "github:kdegroup/ui-agile", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/KDEGroup/UI-AGILE"}], "name": "UI-AGILE", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI agent trained by RL with a continuous reward, a simple-thinking reward and cropping-based resampling on about 9k samples, plus a plug-in decomposed grounding-and-selection inference method.", "topics": ["grounding", "reinforcement-learning"], "updated_at": null, "url": "https://github.com/KDEGroup/UI-AGILE"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["predicted target coordinates"], "counts": [{"basis": "computed", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "population": "sum of icon/text counts in the published card table", "qualifier": "exact", "unit": "target annotations", "value": 1581}], "evaluation": "For positive targets, the predicted point must fall inside the annotated bounding box.", "evidence_urls": ["https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/readme.md", "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py", "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"], "focus": ["GUI grounding"], "notes": "Language, instruction style, and positive/negative target configuration are explicit evaluator options. This profile describes target localization, not end-to-end workflow completion. The target count is computed from the card table, not from expanded language/style evaluation instances.", "observations": ["high-resolution screenshots", "target instructions"], "setting": "offline"}, "leaderboard_url": "https://gui-agent.github.io/grounding-leaderboard"}, "environments": ["desktop", "windows", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "7b64ccad9cfec24c2c9cb35f55923bff93d53729cef3cc7463c79b29bdb1de1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:38:05Z", "role": "primary", "sha256": "a77c859775b33591b5619e8520c94702a53a1f9a29ada93e66744dbe23afe9dc", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["details.benchmark_protocol", "environments"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "c78681e624c748b7a9980b6a9fd1514cec5140c7b3c7f08fe21b36bc7468f2ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/LICENSE"}], "id": "github:likaixin2000/screenspot-pro-gui-grounding", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding"}], "name": "ScreenSpot Pro", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI grounding benchmark focused on professional high-resolution interfaces.", "topics": ["evaluation", "grounding"], "updated_at": null, "url": "https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/LICENSE"}], "id": "github:lll6gg/ui-r1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/lll6gg/UI-R1"}], "name": "UI-R1", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Rule-based reinforcement learning for GUI action prediction, releasing UI-R1-3B, trained from Qwen2.5-VL-3B, the efficiency-oriented UI-R1-E-3B, and training data.", "topics": ["reinforcement-learning", "grounding"], "updated_at": null, "url": "https://github.com/lll6gg/UI-R1"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dafdebf2fa1e0a5db871addd103e50cb55625f0108c757708029e885920ee4e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "98a96c38494df4963951e32b81ec4effebe4c8a812ab8d9e7d185d59e86fe2fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/LICENSE"}], "id": "github:microsoft/gui-actor", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/GUI-Actor"}], "name": "GUI-Actor", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md", "target_id": "arxiv:2506.03143", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding model with attention-based action localization.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/microsoft/GUI-Actor"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:18:21Z", "role": "primary", "sha256": "8aaf8feff6e1d5a533c01afd8afc0cfc0700f0c1c73ba47949b56c11fb8ab5be", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T17:18:56Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/LICENSE"}], "id": "github:microsoft/magma", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T17:18:21Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T17:18:56Z", "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/Magma"}], "name": "Magma", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Microsoft foundation model for multimodal agents, pretrained with Set-of-Mark and Trace-of-Mark objectives on UI, robotics and video data and fine-tunable for web and mobile UI navigation.", "topics": ["grounding", "planning", "mobile-navigation"], "updated_at": null, "url": "https://github.com/microsoft/Magma"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "f4ebaafcc0ab0b9e9cc55cb587678e5c83a5bbd64469fb2cc52bc63843c5cc6d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md"}], "id": "github:microsoft/omniparser", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/OmniParser"}], "name": "OmniParser", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md", "target_id": "arxiv:2408.00203", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screenshot parser that detects interactable regions and captions icons so vision-language models can ground actions, released as V1.5 and V2 weights alongside OmniTool for driving a Windows 11 VM.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://github.com/microsoft/OmniParser"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a0609394a736dc495d077cd57a2945b018123e5b3f9514d7d5bafb9e690aeae2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/LICENSE"}], "id": "github:microsoft/phi-ground", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/Phi-Ground"}], "name": "Phi-Ground", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md", "target_id": "arxiv:2507.23779", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Microsoft GUI grounding models from an empirical study of grounding training, from data collection to model training, released as Phi-Ground-4B-7C with evaluation code.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/microsoft/Phi-Ground"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["pixel (x,y) cursor coordinate"], "counts": [], "evaluation": "The inspected evaluator scales target boxes from a 1000×1000 reference grid into image pixels and tests pixel predictions, allowing ±10 pixels around the center along any target dimension narrower than 1 pixel after scaling. It reports cumulative hit accuracy by turn (earlier hits remain correct), miss distances, any-turn hit rate, and a correction rate whose denominator is all successfully processed samples. The default budget is five turns; retain prompt/feedback variant, turn budget, and processed/skipped coverage.", "evidence_urls": ["https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/README.md", "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/eval/test_allturns_api.py", "https://api.github.com/repos/microsoft/precision-cua-bench/git/trees/main?recursive=1"], "focus": ["pixel-precise editor cursor placement", "multi-turn visual self-correction"], "notes": "The release has four editor/theme JSONL partitions (Cursor/VS Code × dark/light) and screenshot files. Partitions are not train/test splits or independent task totals. README’s initial field table calls bbox pixel coordinates, but the evaluation-format section and actual ratio_h/ratio_w=dimension/1000 code use normalized target boxes; profile follows the inspected evaluator and retains the inconsistency. Source/images were not executed or downloaded as a dataset. MIT is recorded for repository code only.", "observations": ["code-editor screenshot", "cursor-location instruction", "red-cross feedback marking previous incorrect prediction"], "setting": "offline"}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "23d0dc8ea0a10c080cf8ae32121eade2785ae923a6b0d853d166c94c195e789b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://github.com/microsoft/precision-cua-bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "84ad4fb798abb776dce813cf4dab296d6e03248d8ae29388e1f06cccd27d1241", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:08Z", "role": "primary", "sha256": "1fda7c3980fe7a7c373478eda6ecfabff227e92c2b1f19c6fe903b408f583797", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/eval/test_allturns_api.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:06Z", "role": "primary", "sha256": "ede84e3dda47f80fada4234caa2423db08e21ff7667a37599c03aae657b2f84b", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/microsoft/precision-cua-bench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/LICENSE"}], "id": "github:microsoft/precision-cua-bench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/precision-cua-bench"}], "name": "Precision CUA Bench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50", "target_id": "arxiv:2604.13019", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Evaluation resources for pixel-precise cursor grounding in code editors.", "topics": ["evaluation", "grounding"], "updated_at": null, "url": "https://github.com/microsoft/precision-cua-bench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "b3fa8b7e24a01da023a5f21023c4853a90388dd7744055fbb5a2888938c68b0c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/LICENSE"}], "id": "github:microsoft/som", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/SoM"}], "name": "Set of Mark", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/README.md", "target_id": "arxiv:2310.11441", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Visual-prompting toolbox that labels image regions for multimodal reference and grounding, including GUI applications.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/microsoft/SoM"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "c4112415bec58fc77f9a1e8fa98f8d602c5d382747579dfda9f866900c55995c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/LICENSE"}], "id": "github:njucckevin/seeclick", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/njucckevin/SeeClick"}], "name": "SeeClick", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md", "target_id": "arxiv:2401.10935", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding model trained to locate visual interface elements.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/njucckevin/SeeClick"}
{"aliases": ["Oculix", "SikuliX continuation"], "authors": [], "details": {"availability": "Public source and official stable release assets.", "license_components_note": "Project source declares MIT; the retained third-party dependency notices describe additional component terms.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"], "reason": "Publisher project-source LICENSE. Third-party dependency notices are separately retained and no blanket MIT assertion is made for the compiled distribution."}}, "release_tag": "v4.0.0", "repository_revision": "02ea8844483a83a2963db8016cd7ad421e15bc91"}, "environments": ["desktop", "windows", "macos", "linux", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "41d32f40bfb960f38b77ebb62a73ae20454be6c52829c9cbf0b8fedd27de2297", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "151ed1aaf544b9afcfd2d8f13d0fa682d8cf8f8a68496af20c3597b082070801", "supports": ["identity"], "url": "https://api.github.com/repos/oculix-org/Oculix"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:48Z", "role": "primary", "sha256": "cba4541e5d89210b570e42cc69c1529ca36e698622c69b4bb0eeffe56339a759", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/oculix-org/Oculix/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "1022eb9591a2f01833e88c2bae2869faeac9b92e535901d2c79e93a714278777", "supports": ["release_status", "details.release_tag", "environments", "notes"], "url": "https://api.github.com/repos/oculix-org/Oculix/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "8994ff2000b0871aca819ecd7233430bb82cea3f0134a555cebbfa9e02e0f4da", "supports": ["identity", "relations", "notes"], "url": "https://raw.githubusercontent.com/oculix-org/SikuliX1/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:13Z", "role": "primary", "sha256": "102b86e9019fe5e936486b3213d2bbd9e80d6c7ad393284b1a4298b0be555ce7", "supports": ["identity"], "url": "https://api.github.com/repos/RaiMan/SikuliX1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:03Z", "role": "primary", "sha256": "cea4b6292d59625e14968ec666faa7ae6e237b510d846a96875bd6a282677f3a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:50Z", "role": "primary", "sha256": "f0e0d8aa91c1c60bac2250688248c85503036aeac37ec02b8a75fe9d50b6ae4c", "supports": ["licenses.code", "notes"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/API/src/main/resources/Settings/THIRD-PARTY.txt"}], "id": "github:oculix-org/oculix", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T11:18:14Z", "licenses": {"code": {"basis": "Publisher project-source LICENSE. Third-party dependency notices are separately retained and no blanket MIT assertion is made for the compiled distribution.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:03Z", "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/oculix-org/Oculix"}], "name": "OculiX (SikuliX continuation)", "notes": "The official stable-release API lists v4.0.0 with platform-specific API/IDE JARs and an MCP server JAR; no binary was downloaded. The default README still advertises 3.0.3, so release-specific compatibility follows the release notes rather than that older badge. The RaiMan/SikuliX1 API resolves to oculix-org/SikuliX1, whose README calls itself a historical mirror and directs current users here; it is not separately counted. MIT describes the project source. Retained dependency notices also include GPL/LGPL, Apache, BSD and MPL declarations; a bundled artifact is not asserted to be uniformly MIT.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/README.md", "target_id": "github:sikuli/sikuli", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "A maintained continuation of SikuliX that locates on-screen targets with image matching and OCR, drives desktop or Android interfaces, and exposes scripting and optional MCP interfaces.", "topics": ["automation", "grounding", "tool-integration"], "updated_at": null, "url": "https://github.com/oculix-org/Oculix"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": [], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md", "population": "reported multi-level benchmark", "qualifier": "more-than", "unit": "tasks", "value": 8000}], "evaluation": "Level-specific evaluation; Efficiency–Quality Area (EQA) combines navigation quality and efficiency.", "evidence_urls": ["https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md"], "focus": ["screen understanding", "grounding", "task automation", "task collaboration"], "notes": "The README documents four levels, but its release notes and TODOs distinguish released L1/L2 evaluation from pending L3/L4 integration. Do not treat every level as the same evaluator or a complete implemented interface.", "observations": ["screenshots"], "setting": "mixed"}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "35db9fca3ed1c1a3513924fefff4ccd224bc70cd7eb4bf8c54fcb9048851988e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md"}], "id": "github:open-compass/mmbench-gui", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/open-compass/MMBench-GUI"}], "name": "MMBench-GUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md", "target_id": "arxiv:2507.19478", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Four-level GUI-agent benchmark across six platforms, from content understanding and element grounding to task automation and collaboration, scored with an efficiency-quality area metric.", "topics": ["evaluation", "grounding"], "updated_at": null, "url": "https://github.com/open-compass/MMBench-GUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2cf79e21b5438bf6a2e6bd450ba8c4d86404560dfb61836f7e300112ce74a59d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/LICENSE"}], "id": "github:openbmb/agentcpm-gui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenBMB/AgentCPM-GUI"}], "name": "AgentCPM-GUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/README.md", "target_id": "arxiv:2506.01391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "8B on-device smartphone agent built on MiniCPM-V that operates Chinese and English apps from screenshots, trained with reinforcement fine-tuning and a compact JSON action space.", "topics": ["grounding", "mobile-navigation"], "updated_at": null, "url": "https://github.com/OpenBMB/AgentCPM-GUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "021cf14ec9482d1dab04dfeaadc3e39cdc405a284991cd540f5eeae7dea47a12", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/LICENSE"}], "id": "github:os-copilot/os-atlas", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Atlas"}], "name": "OS-Atlas", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/README.md", "target_id": "arxiv:2410.23218", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "OS-Atlas GUI action models, including grounding bases fine-tuned from InternVL2-4B and Qwen2-VL-7B-Instruct that accept any image size and output 0-1000 relative coordinates.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/OS-Copilot/OS-Atlas"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "673b0e309540e57ec3b977689b33f89ef80e45ba1c0ffb05e7412e4862f6b99e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SeeAct/HEAD/README.md"}], "id": "github:osu-nlp-group/seeact", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/SeeAct"}], "name": "SeeAct", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/SeeAct/HEAD/README.md", "target_id": "arxiv:2401.01614", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Generalist web agent framework for large multimodal models such as GPT-4V, with a Playwright tool that runs agents on live websites, text-choice and set-of-mark grounding, and a crawler mode.", "topics": ["web-navigation", "grounding"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/SeeAct"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "72ca3f23114c3a08fefb724573acf10c9f0a1b3d4a50d811c98356264ecf6ffc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/LICENSE.txt"}], "id": "github:osu-nlp-group/uground", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/LICENSE.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/UGround"}], "name": "UGround", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md", "target_id": "arxiv:2410.05243", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Code, training data and weights for universal visual grounding of GUI elements from screenshots, including Qwen2-VL-based UGround-V1 at 2B, 7B and 72B and offline and online evaluation suites.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/UGround"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "8a7f7199f7cf52cb31940bcea8ae797d4ab6dbd2029c3fc9066a17dde3d19095", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "7104ee457c53f2e8e12fd536850e9c2e4894e5e23100f0aa74c8fa5924cec6b2", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/LICENSE"}], "id": "github:othersideai/self-operating-computer", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OthersideAI/self-operating-computer"}], "name": "Self Operating Computer", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Framework, released in November 2023, in which a multimodal model views the screen and picks mouse and keyboard actions toward a goal, with backends such as GPT-4o, Claude 3, Gemini and LLaVA.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/OthersideAI/self-operating-computer"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:38.123692Z", "role": "primary", "sha256": "3226df7608039522072e59d459094fc45be9c5a6be7ea0d7ad76ddf9723a3aea", "supports": ["name", "kind", "summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:04:53.470789+00:00", "role": "primary", "sha256": "3d76d934e98641746085772b98ee17059749f980a616435498a7ad483bd4cec1", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-grounding+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:51:05.389053Z", "role": "primary", "sha256": "4fe6f65a21586dde0bf8b64f8c49e5c583b320bf5e0c2b8253baf2cd3fcfc5d2", "supports": ["id", "url", "notes"], "url": "https://github.com/PatrickStarL/gui-grounding-eval"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:32.022035Z", "role": "primary", "sha256": "c3eb1708b7d57dcde7478ca2f5063241b1ab567b2ba2ba28b9956b043da2b6c2", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/metrics.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:35.461491Z", "role": "primary", "sha256": "66fa5940f0c4f6265cb162cb0596402848dd2a49fd48b07ee950dc866b8730a5", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/geometry.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:38.931130Z", "role": "primary", "sha256": "f83bc0632a616e7b01c8a46b0e435733f0cc47b4a19f0c709b221eaee8a0db5f", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/runner.py"}], "id": "github:patrickstarl/gui-grounding-eval", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-20T11:51:38.123692Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/PatrickStarL/gui-grounding-eval"}], "name": "GUI Grounding Evaluation Harness", "notes": "This is an independent harness using ScreenSpot-Pro, not a new dataset or verified leaderboard entry. The read code maps points back to original-image coordinates and scores point-in-box hits; documentation limits comparisons by prompt, model, subset and provider behavior. No OS support, code license or external data/weight license was established in this bounded review. Model calls, tests and reported results were not executed or reproduced.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "ScreenSpot-Pro evaluation harness with explicit coordinate transforms, prompt and crop-based inference comparisons, and per-sample grounding accuracy and cost records.", "topics": ["grounding", "evaluation", "efficiency"], "updated_at": null, "url": "https://github.com/PatrickStarL/gui-grounding-eval"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "a495bb4c86cf54cd5e2276433cfecb401fd133287406b1a7e04c3eaaac6305ac", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/LICENSE"}], "id": "github:prentisai/groundgui", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:46:05Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/PrentisAI/GroundGUI"}], "name": "GroundGUI training and evaluation code", "notes": "The repository documents separate training-stage environments. Code, model weights, and ScreenRef data have distinct license declarations; no training or benchmark execution was reproduced here.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Code for screenshot grounding through supervised training, teacher-guided distillation, reinforcement learning, and benchmark evaluation.", "topics": ["grounding", "reinforcement-learning", "data-generation"], "updated_at": null, "url": "https://github.com/PrentisAI/GroundGUI"}
{"aliases": [], "authors": [], "details": {"availability": "Public browser-library source and npm package metadata.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"], "reason": "Publisher source license for WebMarker."}}, "package_version": "0.7.0", "repository_revision": "03cf632eef3f76bfbc4fd9c23b31baa0355e41c7"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "b46edb85d41e033f84bdd9bb96122f801364448f28a36fa6deccc110a19b6ffc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "a44253aa866acd38fc3e2275aae61ecd051f956b18cb7b29346c699cf6241d6b", "supports": ["identity"], "url": "https://api.github.com/repos/reidbarber/webmarker"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "fc8de18814136ca3684749363727f263908c23e6ae274a0ed589923774aa3a95", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/reidbarber/webmarker/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "1caae04ee16fcc72c48439677f5aba55128879cd99d6d55fd72caa85a7030e69", "supports": ["identity", "release_status", "details.package_version"], "url": "https://registry.npmjs.org/webmarker-js/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "62dcdb525872e3f49d90ff9b354b79341ac2574522bca7c7146f79040491a627", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"}], "id": "github:reidbarber/webmarker", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T12:30:56Z", "licenses": {"code": {"basis": "Publisher source license for WebMarker.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/reidbarber/webmarker"}], "name": "WebMarker", "notes": "Source and npm release metadata for webmarker-js 0.7.0 are public. The documented mark/unmark API supports screenshot prompts and Playwright interaction, but this utility is not itself a language model or autonomous agent. The WebVoyager/Set-of-Mark references are usage inspiration and are not asserted as papers introducing this library. No package was installed or executed.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A browser library that labels interactive elements and their boxes for visual prompting, preserving element mappings that automation code can use to execute selected actions.", "topics": ["grounding", "screen-understanding", "tool-integration"], "updated_at": null, "url": "https://github.com/reidbarber/webmarker"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/LICENSE"}], "id": "github:ritzz-ai/gui-r1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ritzz-ai/GUI-R1"}], "name": "GUI-R1", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI action-model training and evaluation code with verifiable rewards, plus linked reinforcement-learning data and checkpoints.", "topics": ["reinforcement-learning", "grounding"], "updated_at": null, "url": "https://github.com/ritzz-ai/GUI-R1"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "730e51422c0ebe00cb2a1a8f4510d9914b1ea735fdb06445fde2b243da80bdd4", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "1b4977f976211c2644742e3a56cf75ecffb4c7c31d1bce456875123ba8bcd342", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/LICENSE"}], "id": "github:showlab/showui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/ShowUI"}], "name": "ShowUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/README.md", "target_id": "arxiv:2411.17465", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Lightweight GUI-agent vision-language-action model from Show Lab and Microsoft, released with UI-guided token selection, training code, the ShowUI-web and ShowUI-desktop datasets and vLLM inference.", "topics": ["grounding", "efficiency"], "updated_at": null, "url": "https://github.com/showlab/ShowUI"}
{"aliases": ["ShowUI-pi", "ShowUI Pi"], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "88df552ccbe80efd7020905d77d526f4f9efe2d09a11fed2714b083988f9dcef", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md"}], "id": "github:showlab/showui-pi", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/showui-pi"}], "name": "ShowUI-π", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Vision-language-action model and training code for GUI dragging, with links to checkpoints and continuous-action evaluation data.", "topics": ["grounding", "automation"], "updated_at": null, "url": "https://github.com/showlab/showui-pi"}
{"aliases": [], "authors": [], "details": {"availability": "Public legacy source and original documentation; maintainers explicitly state it is no longer supported.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"], "reason": "Original Sikuli copyright/license file for project source; bundled third-party libraries are not relicensed by this assertion."}}, "maintenance_status": "unsupported legacy research code", "repository_revision": "4adaab7880d2f3e14702ca7287ae9c9e4f4de9ab"}, "environments": ["desktop", "macos", "linux", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:12Z", "role": "primary", "sha256": "e752b2226af679a778e21506e660742376be3d7effe621f63c0b8816d3323729", "supports": ["name", "kind", "summary", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:12Z", "role": "primary", "sha256": "90d840ab98d35b90c0f91f93dc2dc752eac631a4bdc299601bffe8856610dd4c", "supports": ["identity"], "url": "https://api.github.com/repos/sikuli/sikuli"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:49Z", "role": "primary", "sha256": "8ee5f1e75659bfbce0d9b22967ff2b51335be5f19a3dc9e645273cc51333153e", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/sikuli/sikuli/git/trees/develop?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:13Z", "role": "primary", "sha256": "ed7659fff11dac079b99ad3c264f7e1e0a97f313a75379067e1c7fccca23ef88", "supports": ["notes", "details.availability", "relations"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "bab2f2dae13092e9c94f5d4b077bc96a6aa7da5257b54089f76ff5c0b0f61402", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/tutorials/helloworld/helloworld-mac.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "28354a06091997d233cc5124958d85b75a5d815677c58b4fbabd492a6b0038d3", "supports": ["environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/build-linux.txt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "b7c66064d08c6a768a4ed93fee0f86f837b41d51e60785d44a59d7a13bd2452a", "supports": ["environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/build-win32.txt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:03Z", "role": "primary", "sha256": "717e5c54ef8201b2307c9828a6bacb7710385e5014adba93e5f1dc444339fdab", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"}], "id": "github:sikuli/sikuli", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T11:28:12Z", "licenses": {"code": {"basis": "Original Sikuli copyright/license file for project source; bundled third-party libraries are not relicensed by this assertion.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:03Z", "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/sikuli/sikuli"}], "name": "Sikuli legacy research implementation", "notes": "This is the retained Sikuli X-1.0rc3 source line, explicitly marked unsupported by its maintainers. Its documented build targets include historical macOS, Linux and Windows setups; present-day build compatibility was not tested. The official documentation cites the 2009 search/automation and 2010 visual-testing papers. Follow-up SikuliX and current OculiX are distinguished from this legacy research implementation. MIT applies to project code; external libraries keep their own terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst", "target_id": "url:a4cdb2ee0d07d3980651", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst", "target_id": "url:6262c757080f8838e62c", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The original Sikuli source distribution combines screenshot targets with a Jython scripting library and IDE for visual GUI automation and testing.", "topics": ["automation", "grounding", "screen-understanding", "history"], "updated_at": null, "url": "https://github.com/sikuli/sikuli"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "5653db27f736c0c77afb63f81ca051bf31f1e1f0579ac77cf00dca4ea751c159", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "8486a10c4393cee1c25392769ddd3b2d6c242d6ec7928e1414efff7dfb2f07ef", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/LICENSE"}], "id": "github:skyworkai/agent-studio", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T10:59:56Z", "licenses": {"code": {"basis": "repository LICENSE header; GNU text does not say -only or -or-later; not a weights/data license", "declared": "AGPL-3.0", "retrieved_at": "2026-10-18T11:03:47Z", "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/skyworkai/agent-studio"}], "name": "AgentStudio", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/README.md", "target_id": "arxiv:2403.17918", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Toolkit and Ubuntu-based environment for building and evaluating generalist computer-use agents through GUI and API interaction.", "topics": ["sandbox", "evaluation", "grounding", "recording"], "updated_at": null, "url": "https://github.com/skyworkai/agent-studio"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "9f851237932688c15d261c016d51ff5ae80c43c5d0b4eaccac0b48d941118d80", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md"}], "id": "github:tencent/points-gui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tencent/POINTS-GUI"}], "name": "POINTS-GUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md", "target_id": "arxiv:2602.06391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "POINTS-GUI-G-8B GUI grounding model, trained from a POINTS-1.5 base without native grounding through a data pipeline that unifies coordinates, filters noise and adds hard cases.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/Tencent/POINTS-GUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f025e4bc9d920ec982525394c69b87043aae87fbe719de4b9e973408a44211e2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c6a77f53ba7933f07b95807e972e50259921344e777e52a6a8bc690373df2d74", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/LICENSE"}], "id": "github:thudm/cogagent", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/CogAgent"}], "name": "CogAgent", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/README.md", "target_id": "arxiv:2312.08914", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Visual language model for GUI understanding and interaction.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://github.com/THUDM/CogAgent"}
{"aliases": [], "authors": [], "details": {"availability": "Public training configurations, preprocessing and inference examples; full experiment-script release incomplete.", "license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/TongUI-agent/TongUI-agent/git/trees/main?recursive=1"], "reason": "No project-level source-license declaration in the inspected tree; upstream dependency licenses do not establish this project’s grant."}}, "repository_revision": "82631b583180cdc870b4774d2495dd94f1558c46"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "f30a18d1a533b8f71f436dae2ee7369f8971422bd76527e069499c5bba911912", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "90c209a9758f7b527def072cfa78ebc87f1b49b436fad57c4c7bb5a6b66ae790", "supports": ["identity"], "url": "https://api.github.com/repos/TongUI-agent/TongUI-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "0c955a89dcda18652b09ce836c51c6ccaa007be6f6afeb48d21d01813cef1ec2", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/TongUI-agent/TongUI-agent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "b51acb70c95d079ccbe1de0a685ad7c8a4b3b5bed11e6162e6f7c2f60d78805e", "supports": ["summary", "details", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"}], "id": "github:tongui-agent/tongui-agent", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/TongUI-agent/TongUI-agent"}], "name": "TongUI training and inference framework", "notes": "The public tree includes training configs, preprocessing, inference and serving examples. The project still marks release of all experiment/evaluation scripts as work in progress. This entry covers the implementation rather than counting its 3B/7B/32B model-size variants as separate new projects. The inspected project tree has no root source license; linked frameworks, data and checkpoints retain their own terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md", "target_id": "url:8c9aaafb918aaf8b4e92", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md", "target_id": "arxiv:2504.12679", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training configurations, preprocessing and inference examples for tutorial-trained GUI models, with a LLaMA-Factory fine-tuning workflow and screenshot-grounding interfaces.", "topics": ["grounding", "imitation-learning", "data-generation"], "updated_at": null, "url": "https://github.com/TongUI-agent/TongUI-agent"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE", "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md"], "reason": "The implementation and LICENSE live under MAI-UI/. A root-level missing license or the follow-up Qwen-UI-Agent documentation should not change the scoped declaration."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T05:29:14Z", "role": "primary", "sha256": "f644a5c0d0eb29b0ff8d8552dd505c5ecedf7c9cd9ab8cf2c617a1a14ecd0c14", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "fd90bc766d7e73889d14e8d3c5c92d384fa86e163222808f6e492509f00a3656", "supports": ["identity"], "url": "https://api.github.com/repos/Tongyi-MAI/MAI-UI"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:46Z", "role": "primary", "sha256": "4340b3d36bc56963fb8173079da8a0f7e91f195994d29fc47a9f7087cb56ffba", "supports": ["release_status"], "url": "https://api.github.com/repos/Tongyi-MAI/MAI-UI/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T05:29:15Z", "role": "primary", "sha256": "bf989559906a7926b77390082e4267ff7b2b494e8f599a9dc5640319eb45c2ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE"}], "id": "github:tongyi-mai/mai-ui", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T05:29:14Z", "licenses": {"code": {"basis": "Retained MAI-UI/LICENSE and MAI-UI README; applies to the inspected implementation, subject to third-party components.", "declared": "Apache-2.0 (MAI-UI subdirectory)", "retrieved_at": "2026-10-17T05:29:15Z", "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tongyi-MAI/MAI-UI"}], "name": "MAI-UI inference and evaluation code", "notes": "The original implementation is in the MAI-UI/ subdirectory, with an Apache-2.0 LICENSE there. The current root README also promotes Qwen-UI-Agent, but that follow-up directory contains only documentation/display assets in the inspected tree. This entry does not claim the follow-up runtime, training system or weights are released.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md", "target_id": "arxiv:2512.22047", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Published MAI-UI grounding and navigation-agent implementations with inference demos and evaluation scripts.", "topics": ["mobile-navigation", "grounding", "tool-integration"], "updated_at": null, "url": "https://github.com/Tongyi-MAI/MAI-UI"}
{"aliases": ["github:eric-ai-lab/screen-point-and-read", "eric-ai-lab/Screen-Point-and-Read", "https://github.com/eric-ai-lab/Screen-Point-and-Read"], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "7567287ceee0ad2d3d5605abb44ac50463c1008f9b6309393695763cf2ce89bb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/eric-ai-lab/Screen-Point-and-Read/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "4748fb7d92f7a3a2a4a93353db96f6a2df547e5ea9ae87afa811c5423e96f363", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UCSB-AI/Screen-Point-and-Read/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:01Z", "role": "primary", "sha256": "868e18717ff76355c6552a5629b15de9fed1add1262fa5801ecabd7a7fb333d4", "supports": ["aliases"], "url": "https://api.github.com/repos/eric-ai-lab/Screen-Point-and-Read"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:02Z", "role": "primary", "sha256": "868e18717ff76355c6552a5629b15de9fed1add1262fa5801ecabd7a7fb333d4", "supports": ["aliases"], "url": "https://api.github.com/repos/UCSB-AI/Screen-Point-and-Read"}], "id": "github:ucsb-ai/screen-point-and-read", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/UCSB-AI/Screen-Point-and-Read/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UCSB-AI/Screen-Point-and-Read"}], "name": "Screen Point-and-Read / Tree-of-Lens", "notes": "The canonical repository now belongs to UCSB-AI. Its README lists additional setup documentation and evaluation scripts as unfinished.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/eric-ai-lab/Screen-Point-and-Read/HEAD/README.md", "target_id": "arxiv:2406.19263", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screen-reading benchmark and Tree-of-Lens implementation for interpreting user-indicated regions and their layout context.", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": null, "url": "https://github.com/UCSB-AI/Screen-Point-and-Read"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4f578ecd5cbba12797586892a9190e480291bd267a180f0ce56847764eaac524", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/readme.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "d1a708ddcd4afc4e40a5d66f2d56835a12d815515ab6a61578e541ed2c0483bb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/LICENSE"}], "id": "github:vivo/dimo-gui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/vivo/DiMo-GUI"}], "name": "DiMo-GUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/readme.md", "target_id": "arxiv:2507.00008", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding research using modality-aware visual reasoning and test-time scaling.", "topics": ["grounding", "efficiency"], "updated_at": null, "url": "https://github.com/vivo/DiMo-GUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "8cd83b12f4e1c1b22c27b6ae3ee5415564e61ff4072349ec512982a548ea6ec3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "9d26ef3ec511840e10eeaaf1e1b211e196c838823baf6bf9033761f27ab349ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/LICENSE"}], "id": "github:web-infra-dev/midscene", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/web-infra-dev/midscene"}], "name": "Midscene.js", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Visual automation toolkit for web, mobile, and desktop interfaces.", "topics": ["automation", "grounding"], "updated_at": null, "url": "https://github.com/web-infra-dev/midscene"}
{"aliases": [], "authors": [], "details": {"availability": "Public WebVLN training/evaluation and environment-wrapper source; external data/models unverified.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md", "https://api.github.com/repos/WebVLN/WebVLN/git/trees/main?recursive=1"], "reason": "README MIT badge is not backed by a project license file or component-scoped statement in the inspected release."}}, "repository_revision": "59aba284f5888fb7730e1ff07a7e682a1bd1968a"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "cb98feab09648856ab30d8eb739a29e837673cc5b4a6642b7ba3a8943ad71053", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "112e67cc9f865b9640da3827e32eaba5a50f2feb8d85c20018dccbcafa6fd3e5", "supports": ["identity"], "url": "https://api.github.com/repos/WebVLN/WebVLN"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "966adf983bae6052c271f7153c30776a29f7272cf55d05e0175713f02d1900fb", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/WebVLN/WebVLN/git/trees/main?recursive=1"}], "id": "github:webvln/webvln", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/WebVLN/WebVLN"}], "name": "WebVLN research implementation", "notes": "The released tree contains navigation models, an environment wrapper, scoring utilities and train/eval scripts. Full WebVLN-v1 data and pretrained files are linked through an external Drive folder; their complete contents and data/weight terms were not verified, so no separate benchmark download or checkpoint is approved here. The README MIT badge has no matching repository license file or explicit component scope; code permissions remain unestablished.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md", "target_id": "arxiv:2312.15820", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training and evaluation code for question-guided website navigation that combines visual observations with web-specific HTML information.", "topics": ["web-navigation", "grounding", "evaluation"], "updated_at": null, "url": "https://github.com/WebVLN/WebVLN"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dd9f36b818c1b7b6bd772ba8575f6c0ebab6d9e0d307edcc2f196f3e4ba502e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md"}], "id": "github:xlang-ai/aguvis", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/aguvis"}], "name": "Aguvis", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md", "target_id": "arxiv:2412.04454", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Vision-based GUI agent models with unified action representations.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://github.com/xlang-ai/aguvis"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["click point or box center", "refusal represented by negative coordinates in the evaluator"], "counts": [], "evaluation": "The inspected Jedi runner maps predictions from resized-image coordinates back to the original image, then checks the predicted center against rectangle/polygon targets or the refusal rule. It reports overall and classification-group accuracy. This is offline grounding, not OSWorld task-execution success.", "evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md", "https://api.github.com/repos/xlang-ai/OSWorld-G/git/trees/main?recursive=1", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_osworld_g_jedi.py", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"], "focus": ["desktop GUI grounding", "infeasible-target refusal"], "notes": "The repository retains original OSWorld-G.json and refined OSWorld-G_refined.json instructions; treat them as alternative benchmark versions, not additional disjoint tasks. Refined instructions reduce requirements beyond grounding. The same repository also contains Jedi training/rendering material, which is excluded from benchmark counts. The standalone generic eval.py entry point is a stub; the named model runner supplies image_size to its geometry helper. No task count or successful execution is inferred from repository file totals. Apache-2.0 is recorded for code only.", "observations": ["screenshot", "instruction"], "setting": "offline"}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "e814851bc3016a2abf59358f99b80afd9c2bd5894fc34baddfb7e4d0aa03bd45", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "48308bba4f26b02cb7b8fbbc433bb31ec81c494dabd3af437ddbb10b8d9ce667", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/xlang-ai/OSWorld-G/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "64f64e0b847e0fa200aa6051e20b8035a436df00e2ef0d769688ab13fa4d3046", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_osworld_g_jedi.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "b85f5685ce6c200ea648471666177859693d7f40cf50c12d08765ae68580de4a", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/LICENSE"}], "id": "github:xlang-ai/osworld-g", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OSWorld-G"}], "name": "OSWorld-G", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md", "target_id": "arxiv:2505.13227", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-use grounding benchmark with original and refined instructions, released with the Jedi synthetic data pipeline, a 4-million-example Jedi dataset and Jedi-3B and Jedi-7B models.", "topics": ["evaluation", "grounding"], "updated_at": null, "url": "https://github.com/xlang-ai/OSWorld-G"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "2d64349b514c43fb9d0260247300412d33cadc2a029d8279a00cb2a12f26a25a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:43Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/LICENSE"}], "id": "github:yuqi-zhou/gui-g1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:43Z", "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Yuqi-Zhou/GUI-G1"}], "name": "GUI-G1", "notes": "The README retains an older release-pending notice alongside a later checkpoint-release announcement. Check the linked model card for current weight availability and terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/README.md", "target_id": "arxiv:2505.15810", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI-grounding training code that modifies reasoning templates, reward functions, and the GRPO objective, with a linked model release.", "topics": ["grounding", "reinforcement-learning"], "updated_at": null, "url": "https://github.com/Yuqi-Zhou/GUI-G1"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f9ecdf208ed4b7845ea4dbd1a7c936c016d11f7ec28f17e71338ab3c2f5b5c0d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-G2/HEAD/README.md"}], "id": "github:zju-real/gui-g2", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-REAL/GUI-G2"}], "name": "GUI-G2", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-G2/HEAD/README.md", "target_id": "arxiv:2507.15846", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding models at 3B and 7B trained with Gaussian rewards for click proximity to element centers and region coverage, with variance scaled to element size.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/ZJU-REAL/GUI-G2"}
{"aliases": [], "authors": [], "details": {"availability": "Public region-voting evaluation and policy-optimization source; model/data dependencies external.", "license_components_note": "Apache-2.0 notices in VLM-R1 subdirectories do not establish a root-project license.", "license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/ZJU-REAL/GUI-RCPO/git/trees/main?recursive=1", "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/LICENSE", "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/src/open-r1-multimodal/LICENSE"], "reason": "Only subdirectory Apache-2.0 notices were established; no blanket project-source declaration is inferred."}}, "repository_revision": "af15ed5fe8b89b0fe5043a3e94f2984c7b126a4b"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "41c235305d28f6ba77ba1c5d1d9298fd6b1bd68ae3c9ea97eb34a8d22d0ba1dc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/zju-real/gui-rcpo/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "f5823de05f4828a8ab367ef13ec3a8338e020f0876e153ac2e944e4eb0e242c0", "supports": ["identity"], "url": "https://api.github.com/repos/zju-real/gui-rcpo"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "98ed8f0fa50699b8a3d0e8a71cd008113009a3072b971b8fadf2b92f68877e1e", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/ZJU-REAL/GUI-RCPO/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["notes", "licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["notes", "licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/src/open-r1-multimodal/LICENSE"}], "id": "github:zju-real/gui-rcpo", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-REAL/GUI-RCPO"}], "name": "GUI-RC / GUI-RCPO", "notes": "The tree contains GUI-RC evaluation code and VLM-R1-based training scripts, with example JSONL-format paths. It requires separately supplied models and benchmark images; no released trained checkpoint is established here. Apache-2.0 files exist inside VLM-R1 subdirectories, but no repository-wide license is present for all newly added code, so the code slot remains unknown and the component notices are retained.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/zju-real/gui-rcpo/HEAD/README.md", "target_id": "arxiv:2508.05615", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI-grounding research code for spatial voting across sampled predictions and test-time policy optimization using region-consistency rewards.", "topics": ["grounding", "reinforcement-learning", "efficiency"], "updated_at": null, "url": "https://github.com/ZJU-REAL/GUI-RCPO"}
{"aliases": [], "authors": [], "details": {"availability": "Public preprocessing/refinement code, offline action-matching evaluation, and multipart annotation archives; original images and a fine-tuned model are separate.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"], "reason": "Repository GNU GPL version 3 license for code; no data or pretrained-weight license inferred."}}, "repository_revision": "b02e8efdbe02ab82169a2c54d724e33b4037330e"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:55.267942+00:00", "role": "primary", "sha256": "77384ff1dfceac74f2e60af56b016a815963481f9e41a29915698777f55e828e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T11:25:27+00:00", "role": "primary", "sha256": "3d57dc43db78620832e41416e4bfb7ae445afe782e5538d85a9784a8f1e5e4eb", "supports": ["identity"], "url": "https://api.github.com/repos/ZrW00/GUIPivot"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:54.310430+00:00", "role": "primary", "sha256": "a4d3ddbb2b2f4a936c9bc1a1c388b9870dbc5d2c61ff6841a1082b884f5c54e9", "supports": ["release_status", "details.repository_revision", "details.availability", "notes"], "url": "https://api.github.com/repos/ZrW00/GUIPivot/git/trees/b02e8efdbe02ab82169a2c54d724e33b4037330e?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:28.611747+00:00", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:36.501726+00:00", "role": "primary", "sha256": "25e5282e7d85bca133fc77e294cd95d90e4ae65b0d8796cd7c49f53a633b4893", "supports": ["summary", "environments", "notes"], "url": "https://arxiv.org/html/2503.00401v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:30.911159+00:00", "role": "primary", "sha256": "9b5b196e065be2ebc05169a68710b8a6e1bd3b6eb663c68bb449234dbb7a45af", "supports": ["notes", "details.availability"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/data/dataset_info.json"}], "id": "github:zrw00/guipivot", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T12:02:55.267942+00:00", "licenses": {"code": {"basis": "Repository GNU GPL version 3 license for code; no data or pretrained-weight license inferred.", "declared": "GPL-3.0", "retrieved_at": "2026-10-18T12:03:28.611747+00:00", "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZrW00/GUIPivot"}], "name": "GUIPivot", "notes": "The repository releases preprocessing/refinement/evaluation code and multipart ShareGPT-format JSON annotations. Evaluation matches recorded single-step action predictions on AndroidControl and AITZ; it does not run live device tasks. The README links a UIBERT training-JSON path that is not a standalone file in this revision; annotations are packaged in data/GUI.z01, data/GUI.z02 and data/GUI.zip. Original UIBERT/OS-Atlas, AITZ and AndroidControl images must be obtained separately. This entry does not claim a released trained checkpoint or transfer the GPL-3.0 code license to third-party datasets and images.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/README.md", "target_id": "arxiv:2503.00401", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Query-inference training tools connecting screenshot grounding with GUI action reasoning, including preprocessing, query refinement and offline single-step evaluation on AndroidControl and AITZ.", "topics": ["grounding", "planning", "imitation-learning"], "updated_at": null, "url": "https://github.com/ZrW00/GUIPivot"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/tree/main", "gated": false, "hub_id": "Tongyi-MAI/MAI-UI-8B"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "a1087d2a0dde6e7ba6d79db546ee6360a5ffd95ffbc3add1478eedec603a4187", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "349e6d5753fa94a47feb4b2fc53ecd6b9188b645a0cffabef92fd1aa0976ecad", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Tongyi-MAI/MAI-UI-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "a1087d2a0dde6e7ba6d79db546ee6360a5ffd95ffbc3add1478eedec603a4187", "supports": ["licenses.weights"], "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}], "id": "url:00b1a3abf9bcb424b264", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B"}], "name": "MAI-UI 8B", "notes": "Four public safetensors shards are present and Apache-2.0 is declared. The card presents family-wide results and device-cloud architecture; those claims are not all assigned to the 8B checkpoint. Other parameter sizes are not duplicated in this proposal batch.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md", "target_id": "arxiv:2512.22047", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A released MAI-UI checkpoint combining mobile GUI actions, agent-user interaction and tool-use-oriented navigation.", "topics": ["grounding", "mobile-navigation", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md", "https://huggingface.co/api/datasets/xlangai/aguvis-stage1", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage1.yaml"], "formats": ["JSON annotations", "ZIP image packages", "multipart TAR image packages"], "modalities": ["GUI screenshots", "grounding annotations"], "notes": "Stage 1 grounding release, Hub revision 5d11f8f5. JSON annotations and ZIP/multipart TAR image packages are publicly listed. The official training configuration groups web, desktop and mobile inputs and selects all rows of its configured inputs; this is a training recipe, not a release sample count or held-out split. The config names omniact.json while the Hub manifest lists omniact_fix.json, so local paths need reconciliation. No sample total is inferred from names such as webui350k. Apache-2.0 is the release card declaration; upstream source rights remain separate.", "splits": ["training"], "supervision": ["GUI grounding supervised fine-tuning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/tree/main", "gated": false, "hub_id": "xlangai/aguvis-stage1"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "38bda302cf741b9d2b2394b53d157b3c19d8a73e3704c01ed2f9a27c374feb77", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "cb3b8ec88d811d946df3cbf38ef555db5a10d1124668765710c380e9b1988b06", "supports": ["identity", "release_status", "details", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/xlangai/aguvis-stage1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dd9f36b818c1b7b6bd772ba8575f6c0ebab6d9e0d307edcc2f196f3e4ba502e", "supports": ["environments", "summary", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:11Z", "role": "primary", "sha256": "f6a9e55c5d48469e2c8f7799d917d25e7a758036f71c7c045fa92c35057d20da", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage1.yaml"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "38bda302cf741b9d2b2394b53d157b3c19d8a73e3704c01ed2f9a27c374feb77", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}], "id": "url:06dd5ac7917c45d48509", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1"}], "name": "Aguvis stage 1 grounding data", "notes": "The manifest contains grounding JSON annotations and packaged images. The card declares Apache-2.0, which is recorded for this release without overriding source-dataset rights. This first-stage grounding collection is distinct from the project’s second-stage trajectory collection. No archive or training sample was downloaded.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md", "target_id": "arxiv:2412.04454", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The first-stage Aguvis training collection, packaging GUI element-grounding annotations and image archives from multiple interface-data sources.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "population": "context-aware episode training release", "qualifier": "approximately", "unit": "instruction-output pairs", "value": 992000}], "evidence_urls": ["https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"], "formats": [], "modalities": ["screen observations", "instructions", "interaction history"], "notes": "The Hub card contains only licensing metadata. The project README supports the episode-training description and rounded count; archive-level quality and exact row counts remain unverified.", "splits": [], "supervision": ["context-aware grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data/tree/main", "gated": false, "hub_id": "Aria-UI/Aria-UI_Context-aware_Data", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "79f91aab71bb8a0ec3ad9859cbc42d70227e01db729ec80dc5690b466b259cad", "supports": ["name", "kind"], "url": "https://huggingface.co/api/datasets/Aria-UI/Aria-UI_Context-aware_Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["name", "summary", "environments", "topics", "scope", "details.dataset_profile", "relations"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "79f91aab71bb8a0ec3ad9859cbc42d70227e01db729ec80dc5690b466b259cad", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Aria-UI/Aria-UI_Context-aware_Data"}], "id": "url:07b65dc937756ddff4a1", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:11:53Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:11:53Z", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data"}], "name": "Aria-UI Context-aware Data", "notes": "The publisher reports approximately 992K instruction-output pairs in its repository. The Hub card contains only a license declaration; schema quality and exact sample counts were not independently verified.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "github:ariaui/aria-ui", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Episode-based mobile GUI grounding data released with Aria-UI for incorporating interaction history into target localization.", "topics": ["grounding", "memory"], "updated_at": null, "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware/tree/main", "gated": false, "hub_id": "Aria-UI/Aria-UI-context-aware", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "183a8de8928b3b71b171a35149621de078aaf80b4ebde377a5dc3b5cd4a04a27", "supports": ["name", "kind", "summary", "environments", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/Aria-UI/Aria-UI-context-aware"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["summary", "environments", "relations", "release_status", "topics"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "183a8de8928b3b71b171a35149621de078aaf80b4ebde377a5dc3b5cd4a04a27", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/Aria-UI/Aria-UI-context-aware"}], "id": "url:085b6811ddd6f797475d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware"}], "name": "Aria-UI context-aware checkpoint", "notes": "The Hub manifest exposes twelve safetensors shards and model code/configuration, but no README card or model-license declaration. The official repository explicitly announces this context-aware release separately from the base model. Code or base-model licenses are not inherited.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The context-aware Aria-UI release that uses interaction context to support grounding in dynamic Android and desktop agent tasks.", "topics": ["grounding", "interaction-history", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "GUIEnv-local SFT data", "qualifier": "approximately", "unit": "region-text QA pairs", "value": 700000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "text", "bounding boxes"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": ["train", "test"], "supervision": ["OCR", "region-text grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIEnv/tree/main", "gated": false, "hub_id": "yiye2023/GUIEnv", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIEnv/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "15813be05dc76aed5f667875188bf9612e74ff2f4699d270dd0d1aeda86767f8", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIEnv"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "4bcf87ecfbbb8e07a01b21415a970c8b53a5283bf6872b657040d3f45c9241f7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIEnv/raw/main/README.md"}], "id": "url:0af07e00658a430cee6a", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIEnv"}], "name": "GUIEnv-local", "notes": "Public JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate. Only GUIEnv-local is marked released; the advertised 10M-example GUIEnv-global pretraining collection is still unchecked and is not counted here.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The released local-supervision portion of GUIEnv, pairing webpage regions and text for OCR and GUI grounding.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/yiye2023/GUIEnv"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/vyokky/GUI-360/tree/main", "gated": false, "hub_id": "vyokky/GUI-360", "languages": [], "modalities": [], "size_category_reported": ["1M<n<10M"]}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "e51af2845c2f31c7af112579b42c3ff9bf5a2c2356e80932e829bc2be1a341bd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/vyokky/GUI-360"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "e51af2845c2f31c7af112579b42c3ff9bf5a2c2356e80932e829bc2be1a341bd", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/vyokky/GUI-360"}], "id": "url:0b590d03a3d5103a79e4", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/vyokky/GUI-360"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/vyokky/GUI-360"}], "name": "GUI-360", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/vyokky/GUI-360", "target_id": "arxiv:2511.04307", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Desktop GUI trajectory data supporting multiple interface understanding tasks.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/vyokky/GUI-360"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/tree/main", "gated": false, "hub_id": "SenseLLM/GUI-Lasagne-L1"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "270b3af0904a2a50d040948100a327458839dffa5083fc74bb030298f5ab3985", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "224d6487de47cc63251c3b5c5c9f6e75e692b5ef0e12f0a7f71000da0b866e6f", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/SenseLLM/GUI-Lasagne-L1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["environments", "summary"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "270b3af0904a2a50d040948100a327458839dffa5083fc74bb030298f5ab3985", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}], "id": "url:0c5659ca9e59d942aea3", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1"}], "name": "GUI-Lasagne L1", "notes": "Public archive shards are present. The dataset card declares Apache-2.0; it does not resolve rights in third-party interface content. The short card supplies no reliable sample count or explicit train/test partition, so these remain unspecified. The dataset card and archive inventory do not establish its platform distribution; the associated model's platform coverage is not inherited. An associated SpiritSight model card still describes dataset release as forthcoming; the independent dataset card and public archive manifest establish this release's availability.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md", "target_id": "arxiv:2503.03196", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The published GUI-Lasagne training-data collection used for SpiritSight, distributed as dated archive shards for visual interface learning.", "topics": ["grounding", "screen-understanding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "rows", "value": 63530}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "validation split", "qualifier": "exact", "unit": "rows", "value": 7944}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "test split", "qualifier": "exact", "unit": "rows", "value": 7938}], "evidence_urls": ["https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"], "formats": ["Parquet"], "modalities": ["images", "text", "bounding boxes", "OCR metadata"], "notes": "Card-reported row counts are not unique-screen counts. The source field is important for the per-image license terms; the compilation does not relicense original images.", "splits": ["train", "validation", "test"], "supervision": ["instruction-to-target boxes"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/agentsea/wave-ui/tree/main", "gated": false, "hub_id": "agentsea/wave-ui", "languages": [], "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"], "reason": "The compilation expressly does not relicense its images and lists differing source terms. No blanket permissive data license is appropriate."}}, "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "76e5e458eb8788f2c2cfddd06cb944327f363ec70120e122ce93d1350cf82e19", "supports": ["name", "kind", "identity", "release_status", "details", "summary"], "url": "https://huggingface.co/api/datasets/agentsea/wave-ui"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "27ec13aa77dc55113e4d1a19c2269070e2d69d2fe8df8e2313721a1e3fc66bd9", "supports": ["details"], "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "63c6b3ce6a3d9c21a505865ceed698df2894bcb292590c59686e9477df867942", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "e42bbcf8d466c977f9b2150feb73d8cbda2ff907c279630ad4730a8b1f5d665c", "supports": ["summary", "topics", "scope"], "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "76e5e458eb8788f2c2cfddd06cb944327f363ec70120e122ce93d1350cf82e19", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/agentsea/wave-ui"}], "id": "url:0dbb55f127da04d95222", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": {"basis": "Publisher compilation license; individual images keep their source-dataset terms and receive no additional rights from this compilation.", "declared": "Wave UI Dataset License; per-image source terms retained", "retrieved_at": "2026-10-17T04:06:57Z", "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/agentsea/wave-ui"}], "name": "Wave UI", "notes": "Train, validation and test Parquet files are publicly listed. The compilation license explicitly preserves each image’s original source terms and grants no extra rights to those images. The card exposes a platform column without listing its values, so operating-system coverage remains unspecified.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Screenshot-instruction records with target boxes and UI metadata such as OCR, source, element type and platform.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/agentsea/wave-ui"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "SeeClick-derived web subset", "qualifier": "more-than", "unit": "screenshots", "value": 270000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "SeeClick-derived web subset", "qualifier": "more-than", "unit": "UI elements", "value": 3000000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "FineWeb-derived subset", "qualifier": "at-least", "unit": "images", "value": 1600000}], "evidence_urls": ["https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md"], "formats": ["JSON annotations", "image archives"], "modalities": ["screenshots", "referring expressions", "bounding boxes"], "notes": "These are subset-specific counts, not a total for the full OS-Atlas release. Raw boxes use 0–1 coordinates; the documented training wrapper scales them to 0–1000.", "splits": [], "supervision": ["element grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-data", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "c2c5f7b3f07d92349ffd8256792599c856ed77c89108c45dbaef48a4699abe20", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:40:46Z", "role": "primary", "sha256": "19f28b4bc0798e28180b968010c0f7c8dfed2dbfb24b122b958476b871d2e4fd", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "c2c5f7b3f07d92349ffd8256792599c856ed77c89108c45dbaef48a4699abe20", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data"}], "id": "url:0fdeb1cf5f49f84820e6", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data"}], "name": "OS-Atlas Data", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data", "target_id": "arxiv:2410.23218", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding pre-training data for OS-Atlas in desktop (Linux, macOS, Windows), mobile and web domains, with element boxes stored as relative left, top, right and bottom ratios.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/tree/main", "gated": false, "hub_id": "Hcompany/Holo4-35B-A3B"}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "4f1aa5006aa9adaeb5ac549e5a5f8ea164352714a00528b1afb8602812d45b44", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "7a41bdbfb4a300302a7d3633f12962505d56ba874a37292afccbd48b1bec8287", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo4-35B-A3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "4f1aa5006aa9adaeb5ac549e5a5f8ea164352714a00528b1afb8602812d45b44", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}], "id": "url:16083a0225a60882c704", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:11:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:11:45Z", "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B"}], "name": "Holo4 35B-A3B BF16", "notes": "The public manifest contains 26 safetensors shards for the BF16 checkpoint and the card declares Apache-2.0. Quantized FP8/NVFP4/GGUF variants are not split into duplicate entries here. The card also contains demos and scores for Holo4-27B; those are not assigned to this 35B-A3B checkpoint. No training corpus or runtime was downloaded.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A Qwen3.6-based computer-use checkpoint designed to request GUI actions, code execution and tool calls through an external agent harness.", "topics": ["grounding", "action-prediction", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo2-4B/tree/main", "gated": false, "hub_id": "Hcompany/Holo2-4B"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "32232445026dc93113beb6b23d3e43da44d6288a67501f8f0023a78a43efb100", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "bb870c78f3075002ae42a7d1f392fda048fa31a58c865290b2a9b4a1137c6dd3", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo2-4B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "32232445026dc93113beb6b23d3e43da44d6288a67501f8f0023a78a43efb100", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}], "id": "url:16a51d9e725ad5a5c629", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo2-4B"}], "name": "Holo2 4B", "notes": "Two public safetensors shards are present and this 4B model card declares Apache-2.0. The same card assigns different terms to the 30B family member, so the 4B declaration is not generalized to the family. Scores and proprietary training data are not treated as independently verified or publicly released.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A Qwen3-VL-based checkpoint for navigation and multi-step GUI actions across web, desktop and mobile interfaces.", "topics": ["grounding", "planning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/Hcompany/Holo2-4B"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["target localization", "identification of impossible requests"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md", "population": "instruction-variant design", "qualifier": "exact", "unit": "instruction types", "value": 4}], "evaluation": "Publisher supplies original, action, description and negative instruction types; no common end-to-end success metric is asserted.", "evidence_urls": ["https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md", "https://huggingface.co/api/datasets/likaixin/ScreenSpot-v2-variants"], "focus": ["instruction-form robustness", "negative grounding requests"], "notes": "Reuses ScreenSpot-v2 imagery. Image-file count is not treated as a count of independent screenshots or tasks.", "observations": ["screenshots", "instructions"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/tree/main", "gated": false, "hub_id": "likaixin/ScreenSpot-v2-variants"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "7fb2bc9499d8a2b37b663c116af49d11bba493e6767d76a5305fad93257bebf2", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "dea98efa75e063eee2d6bfb5d46da94b41df3d779e2fb0bdaea7866534b5c469", "supports": ["identity", "release_status", "details", "environments"], "url": "https://huggingface.co/api/datasets/likaixin/ScreenSpot-v2-variants"}], "id": "url:17729e1c532ba7043404", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants"}], "name": "ScreenSpot-v2 instruction variants", "notes": "Public image files and desktop/mobile/web annotation JSON files are listed. This is a changed instruction set over an existing screenshot collection, not an independent new screenshot corpus. No data license or precise example count is declared in the card; no paper relationship is inferred from its generic project/PDF links.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Instruction variants over ScreenSpot-v2 screenshots that distinguish original commands, explicit actions, target descriptions and impossible operations.", "topics": ["grounding", "evaluation", "robustness"], "updated_at": null, "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen2.5-VL-3B-Instruct", "gated": false, "hub_revision": "514c49e398fed1203e5c4192be9860131681299e", "license_name": "qwen-research", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "399895cd711d703ff4e016417ca14d13b76988c5625ff120e7083991eb0bd8f2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "c580cb5bd811285944337c4ebb461fa498bba09e6082b70e120ceb83e302fcce", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen2.5-VL-3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "399895cd711d703ff4e016417ca14d13b76988c5625ff120e7083991eb0bd8f2", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}], "id": "url:23ad240757185c6f7eac", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B"}], "name": "Seg2Click Qwen2.5-VL 3B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference. The card names the Qwen Research license, unlike the Apache-2.0 declarations on the Qwen3.5 variants.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Seg2Click checkpoint based on Qwen2.5-VL-3B with a custom action-mask decoder.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-4B", "gated": false, "hub_revision": "318dd331a8a2f60327dfd01e0ffc27998b460a3c", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "98409c81aeba5c9261075ab720204842465041f7df3aee833b28a8e5d4347c56", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "8c94e2115879f474e122cffafe4d14a4e898d832fe094d7f0fffcdba4a14f9d9", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-4B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "98409c81aeba5c9261075ab720204842465041f7df3aee833b28a8e5d4347c56", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}], "id": "url:29588aa5651b5839c045", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B"}], "name": "Seg2Click Qwen3.5 4B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Seg2Click checkpoint pairing a 4B Qwen3.5 backbone with pixel-level action-mask prediction.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK", "population": "VideoCAD published collection", "qualifier": "more-than", "unit": "video recordings", "value": 41000}], "evidence_urls": ["https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK", "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"], "formats": ["ZIP", "MP4", "JSON", "PNG"], "modalities": ["video", "UI action logs", "target CAD images"], "notes": "The reported video total is distinct from the Dataverse package-file count. The exact-version paper identifies the recorded interface as browser-based Onshape; operating-system compatibility is not established.", "splits": [], "supervision": ["timestamped UI actions", "target model images"]}, "dataverse": {"file_count": 261, "license": {"iconUri": "https://licensebuttons.net/p/zero/1.0/88x31.png", "languageCode": "en", "name": "CC0 1.0", "rightsIdentifier": "CC0-1.0", "rightsIdentifierScheme": "SPDX", "schemeUri": "https://spdx.org/licenses/", "uri": "http://creativecommons.org/publicdomain/zero/1.0"}, "persistent_url": "https://doi.org/10.7910/DVN/WX8PCK", "publication_date": "2025-05-15", "publisher": "Harvard Dataverse", "released_at": "2026-03-11T01:35:02Z", "version_minor_number": 0, "version_number": 5, "version_state": "RELEASED"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "f67df0a4422a0484aaa2f93cd185b7000070d2a88d919d9ef995587f5e724007", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:53.536297+00:00", "role": "primary", "sha256": "cbb08803922b4b7736979ca0a5fe552b111fcd428fa9317401bc090b14bb4e54", "supports": ["summary", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:34.311949+00:00", "role": "primary", "sha256": "a303d95a811331cd6ee23b0340f1b9ef2dd996835bfbc66656a871ff94eead50", "supports": ["environments", "notes"], "url": "https://arxiv.org/html/2505.24838v2"}, {"method": "dataverse_api", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "f67df0a4422a0484aaa2f93cd185b7000070d2a88d919d9ef995587f5e724007", "supports": ["details.dataverse", "licenses.data"], "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}], "id": "url:2b5dead3172c0b70da7e", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:20Z", "licenses": {"code": null, "data": {"basis": "publisher Dataverse dataset-version license metadata; original declaration and URI retained in details.dataverse.license", "declared": "CC0 1.0", "retrieved_at": "2026-10-17T02:03:20Z", "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WX8PCK"}], "name": "VideoCAD data", "notes": "Dataverse reports a RELEASED version 5.0 with 261 unrestricted files and a CC0-1.0 dataset declaration. Archive count is not a video or trajectory count. The author repository documents the action and video formats, and Section 3 of the v2 paper identifies the recorded Onshape browser interface. No archives or training samples were downloaded; operating-system coverage is unspecified.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md", "target_id": "arxiv:2505.24838", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "CAD screen recordings paired with action logs and target model images for studying precise interface control and long-horizon visual imitation.", "topics": ["grounding", "imitation-learning", "video"], "updated_at": null, "url": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WX8PCK"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md", "population": "paper-described benchmark coverage", "qualifier": "exact", "unit": "software applications", "value": 83}], "evidence_urls": ["https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md"], "formats": ["JSON annotations", "image files"], "modalities": ["screenshots", "bounding boxes", "UI labels", "action annotations"], "notes": "This is an offline evaluation dataset. Application coverage is not a sample count. The inspected file-layout documentation describes grounding annotations and images; it does not establish a common split or complete action-trajectory inventory.", "splits": [], "supervision": ["element grounding", "layout grounding", "action prediction"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/ServiceNow/ui-vision/tree/main", "gated": false, "hub_id": "ServiceNow/ui-vision", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "34f4ad5c9e7160b5be5aa8ae77e6e9f83a277aada41a272b82a2628d21159a8c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:37Z", "role": "primary", "sha256": "295cf7f7500a18cea5c80eab1cdf4d7264f2f0e959b75efc5ba59083d93c9097", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "34f4ad5c9e7160b5be5aa8ae77e6e9f83a277aada41a272b82a2628d21159a8c", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision"}], "id": "url:2c2004baae4669f8169a", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/ServiceNow/ui-vision"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/ServiceNow/ui-vision"}], "name": "UI-Vision", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision", "target_id": "arxiv:2503.15661", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Desktop GUI data for interface understanding and interaction research.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/ServiceNow/ui-vision"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-Base-4B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "20c6679849cd81a871809aa7f96145855739c9170608dfa6ec3653e52bd795cb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "20c6679849cd81a871809aa7f96145855739c9170608dfa6ec3653e52bd795cb", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B"}], "id": "url:30c831a1198b9bc2a0f5", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B"}], "name": "OS-Atlas Base 4B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B", "target_id": "github:os-copilot/os-atlas", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "OS-Atlas base checkpoint with about 4.1B parameters, fine-tuned from InternVL2-4B and loaded in Transformers through custom InternVL chat model code.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B"}
{"aliases": ["url:cab7bd78abb1faa469e7", "Reallm-Labs/InfiGUI-R1-3B", "https://huggingface.co/Reallm-Labs/InfiGUI-R1-3B"], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/tree/main", "gated": false, "hub_id": "InfiX-ai/InfiGUI-R1-3B"}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "73496fb30838ba9ea291bb6d1e14004ac3318ac03a1ae7b5d6044e0dc14fd46f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-R1-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["identity"], "url": "https://huggingface.co/api/models/Reallm-Labs/InfiGUI-R1-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "5fd1399c88bdf406e7689765ccf0599b3b7e0ddbed082fa92bffdd7b120bd41a", "supports": ["environments"], "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "73496fb30838ba9ea291bb6d1e14004ac3318ac03a1ae7b5d6044e0dc14fd46f", "supports": ["licenses.weights"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:15Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["aliases"], "url": "https://huggingface.co/api/models/Reallm-Labs/InfiGUI-R1-3B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:16Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["aliases"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-R1-3B"}], "id": "url:35b9879a351cc3b3d7d5", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:05:41Z", "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B"}], "name": "InfiGUI-R1 3B", "notes": "The official Hub API resolves the former Reallm-Labs model namespace to InfiX-ai. Two safetensors shards and model/tokenizer files are public and ungated. The card declares Apache-2.0 for this checkpoint; base-model terms remain separate. Its older loading example still uses the prior namespace.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md", "target_id": "arxiv:2504.14239", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL checkpoint trained with the Actor2Reasoner procedure to combine interface grounding with planning and recovery reasoning.", "topics": ["grounding", "reinforcement-learning", "self-reflection"], "updated_at": null, "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md"], "formats": ["Parquet", "LeRobot metadata"], "modalities": ["screenshots", "instructions", "actions"], "notes": "No sample or episode count is inferred from the size tag or package manifest. Main-branch files were inspected; the card’s v3.0 usage example was not executed.", "splits": [], "supervision": ["drag trajectories", "click episodes"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/tree/main", "gated": false, "hub_id": "showlab/ShowUI-pi-data"}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "0e3a2a923865ae24d0de590143e9e32564b0e20db059579582d6181cdcf75bf1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "86fadf9a32185f018f6a8e49d52b39ee5c99c177075510361202dc25a214191f", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-pi-data"}], "id": "url:3dab53c6c38c9e7217ea", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data"}], "name": "ShowUI-π training data", "notes": "Public Parquet chunks and LeRobot metadata are present. The card names a v3.0 revision in its usage example; this review verifies the main manifest and does not assert that tag’s compatibility. No data license is declared. Hub size-category bands are not treated as episode counts.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Drag-and-click episode data for the ShowUI-π model, spanning desktop editing, presentation, handwriting and other interface tasks.", "topics": ["grounding", "action-prediction", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/tree/main", "gated": false, "hub_id": "KDEGroup/UI-AGILE-Data"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "ce35413bc1ef28338bc6fd9f132fa271a4f5e0bf23cecd7346f54adf0a964fd7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "7782422de88830a497d9eb530bd2f0fb6283904cbb25dafcf2ff4d03358a377a", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/KDEGroup/UI-AGILE-Data"}], "id": "url:41abb1b7206689cf5057", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:43:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data"}], "name": "UI-AGILE-Data", "notes": "Public JSON, image archives and Parquet evaluation files are present. The card has no data-license declaration; the model’s MIT declaration is not inherited. The publisher describes fixes to AndroidControl action parameters, but those corrections were not independently reproduced or applied to upstream resources.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "UI-AGILE training annotations and screenshots together with converted GUI-grounding and AndroidControl evaluation files.", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data"}
{"aliases": ["OpenCUA 72B Preview", "OpenCUA-72B-preview", "url:8c792416d297293422cf", "xlangai/OpenCUA-72B-preview", "https://huggingface.co/xlangai/OpenCUA-72B-preview"], "authors": [], "details": {"availability": "Public ungated checkpoint with 82 safetensors shards and model index.", "base_model": "Qwen/Qwen2.5-VL-72B-Instruct", "disabled": false, "download_url": "https://huggingface.co/xlangai/OpenCUA-72B/tree/main", "gated": false, "hub_id": "xlangai/OpenCUA-72B", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"], "reason": "Publisher model card explicitly declares MIT for the released model; other code and training-data scopes are not generalized."}}, "release_revision": "3cc4af5337a1d0010f934284df3c482797266547"}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "ebe551af1008a2307d16cf68715fc1861f8c501c4981660dbe08f691cf73d787", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details", "licenses.weights"], "url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:35Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["identity", "notes"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B-preview"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:04Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["aliases"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B-preview"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:05Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["aliases"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B"}], "id": "url:4a0b1692589594745a3c", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T09:14:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model card explicitly declares MIT for the released model; other code and training-data scopes are not generalized.", "declared": "mit", "retrieved_at": "2026-10-17T09:14:12Z", "url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xlangai/OpenCUA-72B"}], "name": "OpenCUA 72B", "notes": "The former OpenCUA-72B-preview API redirects to the canonical xlangai/OpenCUA-72B repository. The public manifest lists 82 safetensors weight shards plus an index and custom model/tokenizer code. The card declares MIT for the released model and describes absolute coordinates after smart resizing. Weight files and remote code were not downloaded or executed.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md", "target_id": "github:xlang-ai/opencua", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based computer-use checkpoint that combines screenshot history and reflective reasoning to predict executable desktop actions.", "topics": ["grounding", "planning", "action-prediction", "self-reflection"], "updated_at": null, "url": "https://huggingface.co/xlangai/OpenCUA-72B"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B/tree/main", "gated": false, "hub_id": "ByteDance-Seed/UI-TARS-1.5-7B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "73b8964fb8ae2c25009dd4621549a3549d357c4c5c766ef7ed441acaef7efbc9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "73b8964fb8ae2c25009dd4621549a3549d357c4c5c766ef7ed441acaef7efbc9", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B"}], "id": "url:4d32ad3682ae2d71c178", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B"}], "name": "UI-TARS 1.5 7B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B", "target_id": "github:bytedance/ui-tars", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "UI-TARS 1.5 GUI-agent checkpoint in the Qwen2.5-VL architecture with about 8.3B parameters stored in FP32, released under Apache-2.0 for English image-text-to-text use in Transformers.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "prompts", "value": 12600}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "screenshots", "value": 11205}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "desktop applications", "value": 87}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "coordinates"], "notes": "Publisher-reported counts cover a selection and reformatting of GroundCUA, not a new disjoint image collection. No independent image-overlap audit was performed.", "splits": ["train"], "supervision": ["instruction-to-click grounding", "target bounding boxes"]}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "34f0025f686d0e48e2baa5244056a4362018aaad77ac54abea7f1f6eb25c4276", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "34f0025f686d0e48e2baa5244056a4362018aaad77ac54abea7f1f6eb25c4276", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}], "id": "url:4e71dc456c4112eed727", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL"}], "name": "GroundGUI RL Dataset", "notes": "The card declares MIT and attributes the source examples to GroundCUA. Its separate claim about UI-Vision overlap is not independently validated by this catalog.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Desktop grounding prompts selected from GroundCUA and formatted for click-based reinforcement learning with target-box rewards.", "topics": ["grounding", "reinforcement-learning", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/tree/main", "gated": false, "hub_id": "SenseLLM/SpiritSight-Agent-8B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "cc2564a3b2011a06f435e16b95cbf1e22072fd39e76fba844657773541586b37", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/SenseLLM/SpiritSight-Agent-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["licenses.weights"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}], "id": "url:52b7de0e43b1afc9c226", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B"}], "name": "SpiritSight-Agent 8B base", "notes": "The Hub manifest exposes four safetensors shards under SpiritSight-Agent-8B-base. The card declares Apache-2.0 and separately links InternVL terms in its model table. This is the published base checkpoint; benchmark-specific fine-tuning is discussed separately. No model code was executed.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md", "target_id": "arxiv:2503.03196", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An InternVL2-based visual GUI checkpoint distributed for grounding and interface navigation, with the publisher recommending task-specific fine-tuning of the base release.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/neulab/UIX-Qwen2/tree/main", "gated": false, "hub_id": "neulab/UIX-Qwen2"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "91f90e329b76c2ded26a62c41f648a73e685116aa732c86c201592e46a941c44", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "19bd4d126f3ccb4cb3049125d384c76964bd3169c051369ac0c7b64cef9f48b4", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/neulab/UIX-Qwen2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "fe0c03fe73529a29788d559103c4aed48dca6ed9ba3bfd5f151f902fa8858b38", "supports": ["identity", "summary"], "url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "91f90e329b76c2ded26a62c41f648a73e685116aa732c86c201592e46a941c44", "supports": ["licenses.weights"], "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}], "id": "url:54997a3c67e54abdf319", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "odc-by", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/neulab/UIX-Qwen2"}], "name": "UIX-Qwen2", "notes": "Four public safetensors shards are present. The model card declares odc-by, a data-oriented license designation; it is preserved as a publisher declaration rather than replaced with the code or base-model license. No independent legal scope or evaluation result is asserted.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md", "target_id": "arxiv:2410.13824", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A MultiUI-trained visual-language checkpoint for text-rich webpage understanding and interface tasks.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://huggingface.co/neulab/UIX-Qwen2"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "ScreenRef image inventory", "qualifier": "exact", "unit": "distinct screenshots", "value": 585981}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 633369}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "refusal subset of full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 22475}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "coordinates"], "notes": "The publisher reports these inventories in the companion annotation card. The refusal subset is included in the full recipe. Screenshot access is gated; no image download or independent corpus recount was performed.", "splits": ["train"], "supervision": ["instruction-to-click grounding", "bounding boxes", "absent-target abstention"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef/tree/main", "gated": "auto", "hub_id": "PrentisAI/ScreenRef", "languages": ["en"], "license_name": "restricted-research-only", "modalities": ["image", "text"], "size_category_reported": ["100K<n<1M"]}, "environments": ["desktop", "web", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "634b479c7d089e50735f6d4a25f78a63f4d0c2a8c1955015e7641b7afd61c317", "supports": ["name", "kind", "details.gated", "details.hub_id", "licenses.data"], "url": "https://huggingface.co/api/datasets/PrentisAI/ScreenRef"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "634b479c7d089e50735f6d4a25f78a63f4d0c2a8c1955015e7641b7afd61c317", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/PrentisAI/ScreenRef"}], "id": "url:56bda04bd8d3a438deb3", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "other", "retrieved_at": "2026-10-17T01:48:51Z", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef"}], "name": "ScreenRef", "notes": "The public Hub API reports gated=auto and a Restricted Research-Only declaration. The companion card distinguishes annotations from screenshots and limits reuse. Access terms were inspected as public metadata; no access agreement was accepted or gated files downloaded.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI screenshots paired with localization instructions and targets, including absent-target examples; image access is gated under publisher terms.", "topics": ["grounding", "data-generation", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "de16b4fd611982cfaa160cb0fbdc1b0f0559b90983e452964527704a6de4e72d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool"}], "id": "url:57f17b7d0c18b2d7c8bf", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool"}], "name": "Claude Computer Use", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool", "target_id": "organization:anthropic", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Claude tool capability for screenshot-based computer interaction.", "topics": ["planning", "grounding"], "updated_at": null, "url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["click target or bounding box"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md", "population": "test split", "qualifier": "exact", "unit": "examples", "value": 1639}], "evaluation": "Assess localization against annotated targets; agentbrowse, humanbrowse and calendars are distinct source buckets.", "evidence_urls": ["https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md"], "focus": ["instruction-conditioned web localization"], "notes": "The card reports more than 100 websites but does not provide a precise website total. Grounding accuracy does not establish task completion.", "observations": ["web screenshots", "instructions"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/Hcompany/WebClick/tree/main", "gated": false, "hub_id": "Hcompany/WebClick"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "958bbf0eb9f61afd6c576673fbe6f7d84e6c4b8a9d4d8c4fc0e1224b5d8da30b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "c9576957a3cb14785bf274691b50f90b04567a442ed60ee8c39f0f73108359f4", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Hcompany/WebClick"}], "id": "url:5e2bd1b3398974b79c9c", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Hcompany/WebClick"}], "name": "WebClick", "notes": "Four public test Parquet shards are present. The dataset card declares Apache-2.0. This is an offline localization test, not an end-to-end web-task success benchmark; reported screenshots and samples are not summed as different populations.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md", "target_id": "arxiv:2506.02865", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A web-interface localization benchmark with screenshot instructions and target boxes, including agent browsing, human browsing and calendar cases.", "topics": ["grounding", "evaluation", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/Hcompany/WebClick"}
{"aliases": ["url:5f4cb613bda51b45d9c7", "inclusionAI/UI-Venus-2-9B", "https://huggingface.co/inclusionAI/UI-Venus-2-9B"], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b/tree/main", "gated": false, "hub_id": "inclusionAI/UI-Venus-2-9B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "0097be2c0e5c3f3c8aace10de835b0d7c1075531cc7823287c3f5488aebaea08", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "0097be2c0e5c3f3c8aace10de835b0d7c1075531cc7823287c3f5488aebaea08", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:02Z", "role": "primary", "sha256": "3ec7929329e8e1b09f31dee4a3e67128d68cddc95220a5944f4cd6102dae327e", "supports": ["aliases"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:03Z", "role": "primary", "sha256": "3ec7929329e8e1b09f31dee4a3e67128d68cddc95220a5944f4cd6102dae327e", "supports": ["aliases"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}], "id": "url:611df37e5f3caa6befd7", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b"}], "name": "UI-Venus 2 9B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b", "target_id": "github:inclusionai/ui-venus", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "UI-Venus 2 GUI-agent checkpoint fine-tuned from Qwen3.5-9B, with about 9.4B BF16 parameters in the Qwen3.5 multimodal architecture, loaded through Transformers for image-text-to-text.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/ritzzai/GUI-R1/tree/main", "gated": false, "hub_id": "ritzzai/GUI-R1"}, "environments": ["desktop", "windows", "linux", "macos", "android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "3eff86951de7c36dbd4af006ab3f1350213bdfea384494e59c8abc007793c979", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/ritzzai/GUI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["environments"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["licenses.weights"], "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}], "id": "url:61a5321db10fc8a312b9", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/ritzzai/GUI-R1"}], "name": "GUI-R1 checkpoint bundle", "notes": "The manifest contains two safetensors shards in GUI-R1-3B and four in GUI-R1-7B. Both are counted as one Hub resource rather than duplicate root URLs. The model card declares MIT; this does not automatically apply to source training data or upstream base models.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A single Hub release containing GUI-R1 3B and 7B action-model checkpoints trained with reinforcement learning across interface platforms.", "topics": ["grounding", "reinforcement-learning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/ritzzai/GUI-R1"}
{"aliases": [], "authors": ["Tsung-Hsiang Chang", "Tom Yeh", "Robert C. Miller"], "details": {"publication": {"authors": ["Tsung-Hsiang Chang", "Tom Yeh", "Robert C. Miller"], "date_basis": "Supplied month precision: 2010-04. Date Issued: April 2010", "date_parts": [2010, 4], "doi": "10.1145/1753326.1753555", "event_dates": null, "evidence_urls": ["https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Distinct visual-testing contribution from the 2009 Sikuli paper. The institutional abstract and published-version citation supply the retained evidence. Exact publication day and operating-system coverage are not established in the accessible institutional abstract.", "pages": "1535-1544", "pdf_url": null, "publisher": null, "series": null, "title": "GUI Testing Using Computer Vision", "type": "conference-paper", "url": "https://doi.org/10.1145/1753326.1753555", "venue": "Proceedings of the 28th International Conference on Human Factors in Computing Systems (CHI '10)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbace77c54d1b4ed5dcf825dcb0f4525c32e33ac885e70312e8bb4a731a94e80", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbace77c54d1b4ed5dcf825dcb0f4525c32e33ac885e70312e8bb4a731a94e80", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"}], "id": "url:6262c757080f8838e62c", "identifiers": {"doi": "10.1145/1753326.1753555"}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/1753326.1753555"}], "name": "GUI Testing Using Computer Vision", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. This is a separate visual-testing contribution from the 2009 Sikuli paper, not another version of the same article. MIT PDF download returned HTTP 429; its complete institutional abstract and citation metadata remain the retained review source.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents visual GUI test scripts that use images to specify actions and expected screen feedback. Recording both input events and screenshots can generate tests from demonstrations, supporting unit, regression, and test-driven workflows.", "topics": ["evaluation", "grounding", "automation"], "updated_at": null, "url": "https://doi.org/10.1145/1753326.1753555"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md", "https://huggingface.co/api/datasets/cckevinn/GUI-Actor-Data", "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/data/data_config.yaml"], "formats": ["JSON annotations", "ZIP image packages", "multipart image archives"], "modalities": ["GUI screenshots", "bounding boxes", "grounding annotations"], "notes": "Public Hub revision 5a0b35fb includes six bbox JSON files and associated image archives. The training configuration names six input paths with sampling_strategy=all; those input paths are not six disjoint datasets or a verified example count. The card’s source-list wording and mismatched hyperlinks remain unresolved, so its approximately 1M screenshots/10M elements headline is not adopted as a release inventory. No data license is established by the card, and the model repository license is not transferred to these data.", "splits": ["training"], "supervision": ["GUI grounding with bounding-box supervision"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/tree/main", "gated": false, "hub_id": "cckevinn/GUI-Actor-Data"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "e42bbcf8d466c977f9b2150feb73d8cbda2ff907c279630ad4730a8b1f5d665c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.dataset_profile"], "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "d6271cff879b1a2a70490cf3dd95765bef44ecfca531d8ef3545f101673b4423", "supports": ["identity", "release_status", "details", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/cckevinn/GUI-Actor-Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dafdebf2fa1e0a5db871addd103e50cb55625f0108c757708029e885920ee4e", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:12Z", "role": "primary", "sha256": "3521c849b29d7fc88430a8e73a8b2aeb2e613d7fa16ca054dd5bf4ec011b36ca", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/data/data_config.yaml"}], "id": "url:66096ad4057a3d3726f5", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data"}], "name": "GUI-Actor data", "notes": "Public JSON annotations and image archives are present. The card has no data-license declaration. Its source list contains mismatched hyperlinks and says six datasets while naming fewer distinct entries; no source-derivation relationships or aggregate count are imported from that list.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md", "target_id": "arxiv:2506.03143", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A processed screenshot and bounding-box collection released for training GUI-Actor visual grounding models.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "single-step portion", "qualifier": "approximately", "unit": "action instructions", "value": 67000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "multi-step portion", "qualifier": "approximately", "unit": "action instructions", "value": 15000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "instructions", "actions"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": ["train", "test"], "supervision": ["single-step actions", "multi-step navigation"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIAct/tree/main", "gated": false, "hub_id": "yiye2023/GUIAct", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIAct/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["web", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "624ccdd6e6c5b29b588a4289d1d94c1a1c0803868bd586175f2c9ebc3bfdd0b6", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIAct"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "98b45ea81164d1e1a1dd82255207053b15cd6c69d922a1c5cf3387ce604d4b74", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIAct/raw/main/README.md"}], "id": "url:6cd150ec23361b4d2e84", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIAct"}], "name": "GUIAct", "notes": "Public JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Web and Android GUI action-instruction data, with single-step, multi-step and smartphone partitions.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/yiye2023/GUIAct"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-0.8B", "gated": false, "hub_revision": "c36a096cb0cb3320099eb8fde82608cadf2248ee", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "8130a68ede5391bd158ce2cc49c36c5f8c8ebf22b897db6f68e27babbbc66965", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "6bd2608de2715e3d76a0c9e9b8fd7540372bfd2f9864db0445d8e548c9afb518", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-0.8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "8130a68ede5391bd158ce2cc49c36c5f8c8ebf22b897db6f68e27babbbc66965", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}], "id": "url:7405463ba56f9a41e79a", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B"}], "name": "Seg2Click Qwen3.5 0.8B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Compact Seg2Click checkpoint with a merged Qwen3.5 backbone and an action-mask decoder.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "14f573af0c4ea3e5922b8831b7f03724936963e293d7c0aaa3fccbb71dbb0cfa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://openai.com/index/computer-using-agent/"}], "id": "url:74eeb7c9fb20674fe6e2", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/computer-using-agent/"}], "name": "OpenAI CUA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://openai.com/index/computer-using-agent/", "target_id": "organization:openai", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-using model introduced with the Operator research preview.", "topics": ["planning", "grounding"], "updated_at": null, "url": "https://openai.com/index/computer-using-agent/"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md", "population": "default/train in the dataset card metadata", "qualifier": "exact", "unit": "training rows", "value": 701864}], "evidence_urls": ["https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"], "formats": [], "modalities": ["screenshots", "instructions", "functionality labels", "bounding boxes"], "notes": "Rows are not unique screenshots. The card documents image-downscaling conventions that affect coordinates; preserve the device and original image-size fields.", "splits": ["train"], "supervision": ["element grounding", "functionality captioning"]}, "download_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/tree/main", "modalities": ["image", "text", "bounding-box"], "sample_count": 701864, "sample_unit": "training rows reported in the dataset card"}, "environments": ["web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "c73b46acc2ff034a443fb8dbde327de1888535267199c9fb1819adf66898f1d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "c73b46acc2ff034a443fb8dbde327de1888535267199c9fb1819adf66898f1d7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}], "id": "url:79287058342289aa5d45", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T10:42:13Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T10:42:13Z", "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k"}], "name": "AutoGUI v1 training data", "notes": "The row count is not a unique-screen count. The data card documents image scaling conventions that affect coordinate interpretation.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md", "target_id": "arxiv:2502.01977", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screenshot-based grounding and captioning samples with UI functionality descriptions, target coordinates, and bounding boxes.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k"}
{"aliases": ["url:f4bcd696942ee04dcc42", "bevaya/ScreenSpot", "https://huggingface.co/datasets/bevaya/ScreenSpot"], "authors": [], "details": {"benchmark_protocol": {"actions": ["predicted click point; a predicted box is reduced to its center"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md", "population": "Hub test split at 0be08781; not unique screenshots", "qualifier": "exact", "unit": "instruction–target examples", "value": 1272}], "evaluation": "The SeeClick evaluator scores whether the predicted point falls inside the target box, with platform and text/icon breakdowns. The Hub packaging uses xyxy boxes, whereas the original evaluator reads xywh and converts it; adapt the loader to the chosen release.", "evidence_urls": ["https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md", "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot", "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/pretrain/screenspot_test.py"], "focus": ["single-step GUI element grounding"], "notes": "Offline, zero-shot evaluation across mobile, desktop, and web. The card explicitly excludes training use in its intended-use description. Public API reports ungated Parquet files and canonical Hub id bevaya/ScreenSpot; the catalog identity is preserved. Apache-2.0 is the Hub data declaration. No runtime or underlying screenshot rights audit was performed.", "observations": ["screenshot", "English instruction"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot/tree/main", "gated": false, "hub_id": "bevaya/ScreenSpot", "languages": ["en"], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "9e2d2a2c076032225911ae1032b548eafa6331ea81a2779632610f0f1c82dfd9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:00Z", "role": "primary", "sha256": "5ab2240c064502f75a6726311e10e7724ab7055347697f883b15a6ee8e7d00a6", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "eb72113eefd4868114f8a8105eba2d3385da9034a5b3a4860661e03ae7dcae15", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/pretrain/screenspot_test.py"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "9e2d2a2c076032225911ae1032b548eafa6331ea81a2779632610f0f1c82dfd9", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:56Z", "role": "primary", "sha256": "af0c4babeff84158ff196b54bf47242923bbf502d2805a1dd4eb52dba44be89c", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/bevaya/ScreenSpot"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:56Z", "role": "primary", "sha256": "af0c4babeff84158ff196b54bf47242923bbf502d2805a1dd4eb52dba44be89c", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}], "id": "url:7abb418096432f56c685", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot"}], "name": "ScreenSpot", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI grounding test set of 1,272 screenshot instructions from iOS, Android, macOS, Windows and web interfaces, each labeled with the target element's bounding box.", "topics": ["grounding", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/tree/main", "gated": false, "hub_id": "LZXzju/UI-R1-3B-Train"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "640cb5908abefb770ea6bb8d12f22e9eca622224fa73e9f5c67a3a95f4f7b8e5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "482d1304ad53736b3e410694cfa427067ff8b45267374964e4a8ff3833afbbd3", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LZXzju/UI-R1-3B-Train"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["environments", "summary"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}], "id": "url:7c568d88a472f914d316", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train"}], "name": "UI-R1 3B training data", "notes": "The public manifest includes train.json and a train_imgs directory. The card has no dataset-license declaration, and the model’s MIT metadata is not inherited. The official repository identifies this as v1 training data drawn from AndroidControl and ScreenSpot; later enlarged training mixtures are distinct.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/raw/main/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The released mobile training subset for the original UI-R1 3B model, with action annotations and individual screenshot files.", "topics": ["data-generation", "grounding", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-9B", "gated": false, "hub_revision": "76b7e9f7ce97675b7d4e1ea6d88a6adc9ced2886", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "feeb30bda7f86e10d87bdc1b0148d0632aba6df1676931a3d6c65cbdf197df3e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "f215590b7a7b510085f7f99a8c236816604d4c11a8f5ae544b4ee75771dd4f82", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-9B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "feeb30bda7f86e10d87bdc1b0148d0632aba6df1676931a3d6c65cbdf197df3e", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}], "id": "url:826f3753905c5c43bf46", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B"}], "name": "Seg2Click Qwen3.5 9B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Seg2Click checkpoint pairing a 9B Qwen3.5 backbone with a decoder for valid GUI click regions.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-Base-7B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "a3355c4f884490efe2dc31ad393b869349c35130cc37370dafa19f4f0f5016b7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "a3355c4f884490efe2dc31ad393b869349c35130cc37370dafa19f4f0f5016b7", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B"}], "id": "url:82a30f45d9e855f423c7", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B"}], "name": "OS-Atlas Base 7B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B", "target_id": "github:os-copilot/os-atlas", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "OS-Atlas base checkpoint fine-tuned from Qwen2-VL-7B-Instruct, using the standard Qwen2-VL architecture in Transformers with about 8.3B BF16 parameters.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e7062cfb4c0948d3a2274843f9f493f34cb40dfd634371408086df6141a0e036", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "id": "url:879d4e3e75e83582bde6", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "name": "Gemini Computer Use", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://ai.google.dev/gemini-api/docs/computer-use", "target_id": "organization:google", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Gemini capability for generating interface actions from screenshots.", "topics": ["planning", "grounding"], "updated_at": null, "url": "https://ai.google.dev/gemini-api/docs/computer-use"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "41ec857dacdb106ad99abbe1872f40a295056522ffaa1b1714b51dffb370f450", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.agentql.com/"}], "id": "url:8896d44370711136fd54", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.agentql.com/"}], "name": "AgentQL", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Natural-language querying and interaction tools for web interfaces.", "topics": ["grounding", "web-navigation"], "updated_at": null, "url": "https://www.agentql.com/"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "rows", "value": 21988}], "evidence_urls": ["https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md"], "formats": ["Parquet", "JSON", "TAR.GZ"], "modalities": ["screenshots", "text", "bounding boxes", "points"], "notes": "Row count is the card dataset_info value; one row may contain multiple instructions. The card summarizes screenshots as approximately 22K. No data license is declared.", "splits": ["train"], "supervision": ["element grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/showlab/ShowUI-web/tree/main", "gated": false, "hub_id": "showlab/ShowUI-web"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "08586aa1ded1f1a1f984fb55697fdb1007d86c7fdf4988a2dc1fcf3fdfd49a47", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "12412480f1a3957069d9dcdc53abce4ad3e34be7dac7ca6fe1726d04764cadfd", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-web"}], "id": "url:8933977bb0a57940ed23", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/showlab/ShowUI-web"}], "name": "ShowUI-web", "notes": "Public Parquet annotations, metadata JSON and images.tar.gz are present. The card refers to image.tar.gz in its extraction command, while the actual archive is plural. No data-license declaration is present; the ShowUI code or model licenses are not inherited.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md", "target_id": "arxiv:2411.17465", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Web-interface grounding examples with instructions, element boxes and points, paired with a separately packaged screenshot archive.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/showlab/ShowUI-web"}
{"aliases": [], "authors": [], "details": {"availability": "Ungated annotation files and multipart image-archive manifest.", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md", "population": "TongUI-collected GUI-Net-1M data, excluding added external mixtures", "qualifier": "approximately", "unit": "training trajectories", "value": 1000000}], "evidence_urls": ["https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md", "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"], "formats": ["JSON annotations", "multipart image archives"], "modalities": ["text", "screenshots"], "notes": "The source reports an approximate trajectory total; it does not establish per-platform counts or a complete test-split inventory.", "splits": ["training collection"], "supervision": ["GUI trajectory supervision from multimodal tutorials"]}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"], "reason": "Publisher declaration for this dataset packaging; original web/tutorial materials retain applicable rights."}}, "release_revision": "60d1e644fcff7d3b62093baec03fe7efeb0d98a1"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "f30a18d1a533b8f71f436dae2ee7369f8971422bd76527e069499c5bba911912", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "314a77eb0ba43bedffbe6869601e41aea41993eaca59d86ba346652c853802d2", "supports": ["identity", "release_status", "details.release_revision", "details.dataset_profile.formats"], "url": "https://huggingface.co/api/datasets/Bofeee5675/GUI-Net-1M"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "b51acb70c95d079ccbe1de0a685ad7c8a4b3b5bed11e6162e6f7c2f60d78805e", "supports": ["summary", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "c2df5c3a6dd576ae53f0dc0f72ecab07d93ed034b04336abd2072a6f0d812c64", "supports": ["licenses.data", "details.dataset_profile.formats", "notes"], "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"}], "id": "url:8c9aaafb918aaf8b4e92", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": {"basis": "Publisher declaration for this dataset packaging; original web/tutorial materials retain applicable rights.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T12:32:07Z", "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M"}], "name": "GUI-Net-1M", "notes": "The ungated manifest lists JSON annotations and multipart image archives. Training documentation describes roughly one million collected training trajectories from Wikihow, Baidu Experience and GUI videos; this is distinct from larger mixtures that add external datasets. Platform-specific distribution and exact row counts were not established from this release metadata. Apache-2.0 is the publisher dataset declaration, without a grant of additional rights to upstream tutorial content. No image archive or example record was downloaded.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md", "target_id": "arxiv:2504.12679", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A TongUI-collected training corpus from multimodal web tutorials, distributed as annotation files and split screenshot archives for GUI trajectory supervision.", "topics": ["data-generation", "grounding", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/openbmb/AgentCPM-GUI/tree/main", "gated": false, "hub_id": "openbmb/AgentCPM-GUI"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "9016a68e833a7717c922d9363a6d6c56cc7cf6a7087543f5694507668fe9ab75", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "4d76c8cc19f1ff4c850ba6cc566d50bbd2355e10921a2214e38c921cedfeed4c", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/openbmb/AgentCPM-GUI"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "9016a68e833a7717c922d9363a6d6c56cc7cf6a7087543f5694507668fe9ab75", "supports": ["licenses.weights"], "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}], "id": "url:8ef3985dce697e5feea9", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/openbmb/AgentCPM-GUI"}], "name": "AgentCPM-GUI 8B checkpoint", "notes": "The Hub provides four safetensors shards and associated model files without gating. Weight metadata declares Apache-2.0, while the prose license section refers specifically to code; these scopes are kept distinct. The card describes Android-app operation, and the CAGUI evaluation data remains a separate existing catalog asset.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md", "target_id": "arxiv:2506.01391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An eight-billion-parameter MiniCPM-V-based model that consumes smartphone screenshots and predicts compact actions for Chinese and English Android app tasks.", "topics": ["mobile-navigation", "grounding", "reinforcement-learning"], "updated_at": null, "url": "https://huggingface.co/openbmb/AgentCPM-GUI"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/inclusionAI/GUI-G2-7B/tree/main", "gated": false, "hub_id": "inclusionAI/GUI-G2-7B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "cf53a868a010a20b165386edaaafed8918071c35a3babf1766fbb51f020cb4a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "c1f11fea30d3a64445e983a011db4efb9515a4221ad770e3a5223699e665eebe", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/inclusionAI/GUI-G2-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "cf53a868a010a20b165386edaaafed8918071c35a3babf1766fbb51f020cb4a1", "supports": ["licenses.weights"], "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}], "id": "url:90eea6f6d099a3297850", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/inclusionAI/GUI-G2-7B"}], "name": "GUI-G2 7B", "notes": "The model card and four public safetensors shards establish this specific release. Apache-2.0 is declared for the model. The release targets element grounding; its grounding results do not establish long-horizon task-completion success.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md", "target_id": "arxiv:2507.15846", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL grounding checkpoint trained with Gaussian point and coverage rewards whose granularity adapts to interface-element scale.", "topics": ["grounding", "reinforcement-learning", "reward-models"], "updated_at": null, "url": "https://huggingface.co/inclusionAI/GUI-G2-7B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/tree/main", "gated": false, "hub_id": "Yuqi-Zhou/GUI-G1-3B-v1"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "b7a4628a4d65ae226b50086075f3977b8f69ad12d15238a194675875bb9113e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "2f4b2084ca9cfd67f0952ea8c578c87f3a5d721064b5de3a191d7489373125aa", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Yuqi-Zhou/GUI-G1-3B-v1"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "b7a4628a4d65ae226b50086075f3977b8f69ad12d15238a194675875bb9113e8", "supports": ["licenses.weights"], "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}], "id": "url:9dc9d8d3b9a4b32d6f4b", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1"}], "name": "GUI-G1 3B v1", "notes": "Two public safetensors shards and model/tokenizer configuration are present. The card declares MIT. Its inference example loads the v1 model but names a different 0.1K processor checkpoint; that example was not executed, and compatibility is not asserted.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md", "target_id": "arxiv:2505.15810", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-3B checkpoint released for R1-style visual grounding experiments, predicting interface target locations from screenshots and instructions.", "topics": ["grounding", "reinforcement-learning"], "updated_at": null, "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/tree/main", "gated": false, "hub_id": "Hcompany/Holo3-35B-A3B"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0681f87b059110a4e5dff8955e9feca6c5ea118b326f729f9c0ed0fe49e42d03", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "bc5324dd6ceea9a0cc6bf0639b99d24cdd172af3c5d247bf923fdce1dc11f582", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo3-35B-A3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0681f87b059110a4e5dff8955e9feca6c5ea118b326f729f9c0ed0fe49e42d03", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}], "id": "url:a349aace3cfde25628a8", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B"}], "name": "Holo3 35B-A3B", "notes": "Fourteen public safetensors shards are present. The card declares Apache-2.0 for this release and describes a 35B-total/3B-active MoE architecture. No unpublished enterprise benchmark or training corpus is added solely from its performance claims.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A sparse Qwen3.5-based GUI-action checkpoint released for navigation and computer-use workflows.", "topics": ["grounding", "planning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B"}
{"aliases": [], "authors": ["Tom Yeh", "Tsung-Hsiang Chang", "Robert C. Miller"], "details": {"publication": {"authors": ["Tom Yeh", "Tsung-Hsiang Chang", "Robert C. Miller"], "date_basis": "Supplied month precision: 2009-10. Date Issued: October 2009", "date_parts": [2009, 10], "doi": "10.1145/1622176.1622213", "event_dates": null, "evidence_urls": ["https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Institutional abstract and published-version citation supply the retained evidence; the institution identifies an author final manuscript record. Exact publication day is not supplied. Operating-system-specific experiments are not asserted.", "pages": "183-192", "pdf_url": null, "publisher": null, "series": null, "title": "Sikuli: Using GUI screenshots for search and automation", "type": "conference-paper", "url": "https://doi.org/10.1145/1622176.1622213", "venue": "Proceedings of the 22nd Annual ACM Symposium on User Interface Software and Technology (UIST '09)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbaaf21982522b34857be014b39d6c43fedcda20f808efe40d04958c77243a5a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbaaf21982522b34857be014b39d6c43fedcda20f808efe40d04958c77243a5a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"}], "id": "url:a4cdb2ee0d07d3980651", "identifiers": {"doi": "10.1145/1622176.1622213"}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/1622176.1622213"}], "name": "Sikuli: Using GUI screenshots for search and automation", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. MIT identifies the retained content as an author final manuscript record. Its PDF download returned HTTP 429; no retry was made. The complete institutional abstract and bibliographic record were read.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces screenshot-based GUI search and visual scripting, using image patterns to locate interface targets and direct mouse and keyboard events. It evaluates screenshot queries and demonstrates automation and interactive-help tasks.", "topics": ["grounding", "automation", "screen-understanding"], "updated_at": null, "url": "https://doi.org/10.1145/1622176.1622213"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset/tree/main", "gated": false, "hub_id": "osunlp/GUI-Drag-dataset", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "cdb8be8c4692a8ddcf8a66c1f0be666a9214718445925fb3c88615a0d641bf3a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/osunlp/GUI-Drag-dataset"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:54Z", "role": "primary", "sha256": "907c8a5169c180491ef8c26c9428b2c67bc4e26f472d6423ce3715cbe47f76ef", "supports": ["notes"], "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "cdb8be8c4692a8ddcf8a66c1f0be666a9214718445925fb3c88615a0d641bf3a", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/osunlp/GUI-Drag-dataset"}], "id": "url:aa28e1329a52e341d285", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset"}], "name": "GUI-Drag", "notes": "The card links its paper (arXiv 2601.06031) and code repository (OSU-NLP-Group/GUI-Drag). Neither is a reviewed catalog record yet, so no relation is recorded.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Data for learning text-dragging interactions in graphical interfaces.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/xlangai/OpenCUA-32B/tree/main", "gated": false, "hub_id": "xlangai/OpenCUA-32B"}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "151217eb96e4ab22a47c98766052e5cfd76497a2a9197cc125c0237f3fbc59f6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "a3be2e1a6cd39ee89809baa9504813c4a6a9a03bbd916e82924cc0f89cb5da33", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-32B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "151217eb96e4ab22a47c98766052e5cfd76497a2a9197cc125c0237f3fbc59f6", "supports": ["licenses.weights"], "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}], "id": "url:ad7f176848fd84de64e2", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xlangai/OpenCUA-32B"}], "name": "OpenCUA 32B", "notes": "The public manifest contains 64 safetensors shards. The card declares MIT for this model and names its Qwen2.5-VL base; upstream terms remain separate. This is the 32B release, distinct from the existing 72B preview. Family-level or 72B benchmark scores are not transferred to it.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based computer-use checkpoint that combines screenshot history with planning and grounded executable actions.", "topics": ["grounding", "planning", "self-reflection"], "updated_at": null, "url": "https://huggingface.co/xlangai/OpenCUA-32B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/mPLUG/GUI-Owl-7B/tree/main", "gated": false, "hub_id": "mPLUG/GUI-Owl-7B"}, "environments": ["android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "a0e6dd76aee6f138c15f244f4e2e8ba650b8829295397fdbfa839a8edf852b57", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "4822eab7fd35e583d52db7369d7d929d69dbde66b27c9da3d370cad9ec06ec89", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/mPLUG/GUI-Owl-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "a0e6dd76aee6f138c15f244f4e2e8ba650b8829295397fdbfa839a8edf852b57", "supports": ["licenses.weights"], "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}], "id": "url:aeb32973694f91ad9920", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/mPLUG/GUI-Owl-7B"}], "name": "GUI-Owl 7B", "notes": "Five public safetensors shards are present. MIT is declared in this model card. This record identifies the original 7B release; other sizes and later GUI-Owl generations are not aliases or assumed to share its exact terms. The online demo is still marked forthcoming.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md", "target_id": "arxiv:2508.15144", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based GUI checkpoint released with Mobile-Agent-v3 for grounded actions and specialized roles in its agent framework.", "topics": ["grounding", "action-prediction", "planning"], "updated_at": null, "url": "https://huggingface.co/mPLUG/GUI-Owl-7B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/Hcompany/DragOn/tree/main", "gated": false, "hub_id": "Hcompany/DragOn", "languages": ["en", "fr"], "modalities": [], "size_category_reported": ["1M<n<10M"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "08b5529ce01781143147d56e2278aec2e5bdd972ba7e2fae6ba7e65e62688d40", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/Hcompany/DragOn"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:54Z", "role": "primary", "sha256": "e7ef94cafcfd155961beec373f2f4e727715b7d517a819f52b5407fc2dec4009", "supports": ["notes"], "url": "https://huggingface.co/datasets/Hcompany/DragOn/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "08b5529ce01781143147d56e2278aec2e5bdd972ba7e2fae6ba7e65e62688d40", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Hcompany/DragOn"}], "id": "url:b854ff53e497da3a5ad8", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/Hcompany/DragOn"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Hcompany/DragOn"}], "name": "DragOn", "notes": "The card cites a workshop paper titled \"DragOn: A Drag-Grounding Benchmark and Training Dataset for GUI Agents\" without an arXiv link. arXiv 2606.06322 has a different title, so no paper relation is recorded.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Drag-grounding training and evaluation data pairing each screenshot instruction with start and end boxes, across text highlighting, spreadsheet cells, slide resizing and sliders.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/datasets/Hcompany/DragOn"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/KDEGroup/UI-AGILE/tree/main", "gated": false, "hub_id": "KDEGroup/UI-AGILE"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "75aed1ebff7dd54ab1dc0112fe524a118d9e8845c2e079080d2bb7954f7ab36e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "a52032bb185885d47466f0bea26c15ccefeb8d6ff89f3bacc5eb4192c35d45f7", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/KDEGroup/UI-AGILE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "ce35413bc1ef28338bc6fd9f132fa271a4f5e0bf23cecd7346f54adf0a964fd7", "supports": ["environments"], "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "75aed1ebff7dd54ab1dc0112fe524a118d9e8845c2e079080d2bb7954f7ab36e", "supports": ["licenses.weights"], "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}], "id": "url:b94e285c3f757e2d0f3a", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/KDEGroup/UI-AGILE"}], "name": "UI-AGILE checkpoint", "notes": "Four safetensors shards and model/tokenizer files are public. The card declares MIT for these weights; the companion dataset has no declared license and is separate. The short card does not state a parameter count, so this entry does not infer one from shard sizes.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The released UI-AGILE model for GUI grounding with reward shaping, crop-based resampling and decomposed selection during inference.", "topics": ["grounding", "reinforcement-learning", "efficiency"], "updated_at": null, "url": "https://huggingface.co/KDEGroup/UI-AGILE"}
{"aliases": [], "authors": ["Toby Jia-Jun Li", "Igor Labutov", "Xiaohan Nancy Li", "Xiaoyi Zhang", "Wenze Shi", "Wanling Ding", "Tom M. Mitchell", "Brad A. Myers"], "details": {"publication": {"authors": ["Toby Jia-Jun Li", "Igor Labutov", "Xiaohan Nancy Li", "Xiaoyi Zhang", "Wenze Shi", "Wanling Ding", "Tom M. Mitchell", "Brad A. Myers"], "date_basis": "Supplied year precision: 2018. APPINITE publication citation: VL/HCC 2018", "date_parts": [2018], "doi": null, "event_dates": null, "evidence_urls": ["https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf", "https://www.cs.cmu.edu/~NatProg/sugilite.html", "https://toby.li/publications/"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The base SUGILITE framework and later PUMICE work are distinct publications; this paper focuses on resolving data-description ambiguity in demonstrations. Author-hosted manuscript and institutional publication lists supply the retained evidence. DOI, final page range, and exact publication day remain unknown.", "pages": null, "pdf_url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf", "publisher": null, "series": null, "title": "APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions", "type": "conference-paper", "url": "https://ieeexplore.ieee.org/document/8506506", "venue": "Proceedings of the 2018 IEEE Symposium on Visual Languages and Human-Centric Computing", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:46Z", "role": "primary", "sha256": "6e05fd08b187adba32148e0835ef6b9e20cd6a3ba5e0de6e2778eb18c02788ee", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title"], "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:20Z", "role": "primary", "sha256": "adfcec73830b612b181d4846ace3fbaab1768e906715c1d90fe657e23fe4d985", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://toby.li/publications/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:46Z", "role": "primary", "sha256": "6e05fd08b187adba32148e0835ef6b9e20cd6a3ba5e0de6e2778eb18c02788ee", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "environments", "name", "scope", "summary", "topics"], "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:20Z", "role": "primary", "sha256": "adfcec73830b612b181d4846ace3fbaab1768e906715c1d90fe657e23fe4d985", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://toby.li/publications/"}], "id": "url:ba2c664c1b3c6f1b172f", "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:54:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://ieeexplore.ieee.org/document/8506506"}, {"rel": "publication_pdf", "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}], "name": "APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The base SUGILITE framework and later PUMICE work are distinct publications; this paper focuses on resolving data-description ambiguity in demonstrations.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces APPINITE, a conversational interface for specifying the intended targets of demonstrated Android actions. It constructs semantic UI representations and uses an on-screen overlay to preview and refine executable data descriptions.", "topics": ["grounding", "human-in-the-loop", "mobile-navigation"], "updated_at": null, "url": "https://ieeexplore.ieee.org/document/8506506"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["click coordinates"], "counts": [], "evaluation": "The inspected Jedi evaluation adapter reads the v2 mobile/desktop/web annotation files, scores a point against the target box, and reports overall, platform, and text/icon accuracy. This is a named compatible evaluator implementation; the OS-Atlas dataset card itself supplies no evaluator specification.", "evidence_urls": ["https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2", "https://osatlas.github.io/", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"], "focus": ["single-step GUI grounding on the ScreenSpot-v2 release"], "notes": "The public Hub revision 5efbb1f1 contains three platform JSON annotation files and an image ZIP; its card only declares Apache-2.0. No sample total or equivalence to the original ScreenSpot test count is asserted. Keep v2 separate from v1 and retain evaluator/version and coordinate conversion when comparing results. Files were listed, not downloaded as data.", "observations": ["screenshot", "instruction"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2/tree/main", "gated": false, "hub_id": "OS-Copilot/ScreenSpot-v2", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "7feddadc306e384f8df6a3d9c8afc0c776cb3734c4c7fe3eb842eab85537f408", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:05Z", "role": "primary", "sha256": "94547cb88756cf6a36bfb86d5fe9bdca1e40d7edb9e452c33cb462bbe760fed0", "supports": ["details.benchmark_protocol"], "url": "https://osatlas.github.io/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "0e8772bc8fd5a25c50d3c2f6b0caee05ed2ab4659d46d99d60adf2b5ab29e997", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "b85f5685ce6c200ea648471666177859693d7f40cf50c12d08765ae68580de4a", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "7feddadc306e384f8df6a3d9c8afc0c776cb3734c4c7fe3eb842eab85537f408", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2"}], "id": "url:bd7214de6d98917c2350", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"}], "name": "ScreenSpot v2", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Second version of the ScreenSpot grounding test set, published by OS-Copilot as separate desktop, mobile and web annotation files with a shared screenshot archive.", "topics": ["grounding", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "84ce0fd682367a6dac5cc5abc39d5c877556024bbf6ffdad30f76fc2ec08523f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://hcompany.ai/research/models"}], "id": "url:c0bd07d82ce25ad1d99d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://hcompany.ai/research/models"}], "name": "Holo Model Family", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://hcompany.ai/research/models", "target_id": "organization:h", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "H computer-use model family and its release-specific access options.", "topics": ["planning", "grounding"], "updated_at": null, "url": "https://hcompany.ai/research/models"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/showlab/ShowUI-2B/tree/main", "gated": false, "hub_id": "showlab/ShowUI-2B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "bb6d120d4795492a197e0a9e9df5c56b7baa8ac5c08182ce0f20f36c8b65783d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/showlab/ShowUI-2B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "bb6d120d4795492a197e0a9e9df5c56b7baa8ac5c08182ce0f20f36c8b65783d", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/showlab/ShowUI-2B"}], "id": "url:c0f39f7cf2cd9c18e8bd", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/showlab/ShowUI-2B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/showlab/ShowUI-2B"}], "name": "ShowUI 2B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/showlab/ShowUI-2B", "target_id": "github:showlab/showui", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "ShowUI vision-language-action checkpoint fine-tuned from Qwen2-VL-2B-Instruct, published as a single PyTorch weight file under the MIT license and tagged for GUI agents and computer use.", "topics": ["grounding", "efficiency"], "updated_at": null, "url": "https://huggingface.co/showlab/ShowUI-2B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 633369}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "point recipe", "qualifier": "exact", "unit": "annotation rows", "value": 610894}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"], "formats": ["JSONL", "JSON task recipes", "TSV manifests"], "modalities": ["text", "coordinates", "image references"], "notes": "The recipes reuse source images and are not additive datasets. This download contains no screenshots. Counts are publisher-reported; rows named for trajectories are independent decision examples, not full trajectories.", "splits": ["full/train", "point/train"], "supervision": ["instruction-to-click grounding", "bounding boxes", "absent-target abstention"]}, "license_name": "restricted-research-only"}, "environments": ["desktop", "web", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}], "id": "url:c1db2ba79e6d6a3db2b9", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations"}], "name": "ScreenRef Annotations", "notes": "The card declares restricted academic research use and no redistribution. Annotation coordinates use a normalized 1000 by 1000 space. Some source instructions name gestures other than the single-click target; the publisher recommends treating those as location supervision.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}, {"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "url:56bda04bd8d3a438deb3", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Annotation files, task recipes, and image manifests for ScreenRef; screenshots reside in the separately gated image repository.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/api/datasets/osunlp/UGround-V1-Data", "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"], "formats": ["Parquet shards"], "modalities": ["GUI screenshots", "natural-language grounding descriptions"], "notes": "Public API revision 33325877 lists Parquet assets but gated=auto; the pinned raw card returned HTTP 401 and no access was requested. Official README identifies this as V1 training data; the API excerpt also announces a Web-Hybrid bounding-box variant with coordinates normalized to [0,999]. That does not establish every shard schema or a release-wide point/bbox count. No train/test partition, example total, or unrestricted downloadability is claimed. Preserve the recorded CC-BY-NC-SA-4.0 data declaration separately from model/code terms.", "splits": [], "supervision": ["GUI visual-grounding training for the UGround-V1 family"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data/tree/main", "gated": "auto", "hub_id": "osunlp/UGround-V1-Data", "languages": [], "modalities": ["tabular", "text"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "0342862bd7857dd21e389ee5d351375dfea722cad4725c8e33271efc93e009f6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/osunlp/UGround-V1-Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "72ca3f23114c3a08fefb724573acf10c9f0a1b3d4a50d811c98356264ecf6ffc", "supports": ["details.dataset_profile", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "0342862bd7857dd21e389ee5d351375dfea722cad4725c8e33271efc93e009f6", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/osunlp/UGround-V1-Data"}], "id": "url:c4898ad1bfc049926a63", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-nc-sa-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data"}], "name": "UGround V1 Data", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md", "target_id": "github:osu-nlp-group/uground", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Gated web grounding training data for UGround in Web-Hybrid and Web-Direct subsets, with a bounding-box version whose coordinates are normalized to 0-999.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data"}
{"aliases": ["url:4a309ef39b33a7d1e543", "THUDM/cogagent-9b-20241220", "https://huggingface.co/THUDM/cogagent-9b-20241220"], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/zai-org/cogagent-9b-20241220/tree/main", "gated": false, "hub_id": "zai-org/cogagent-9b-20241220"}, "environments": ["desktop", "windows", "macos", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "de2c8e7a9792c3d129451ec813d458f83263460a5808cf87c45b04b260a376c9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:05Z", "role": "primary", "sha256": "ae72fbc40f9c8afd52d1688b2233e5c915571de33618cfd1ffcb7f3fadfb4402", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/cogagent-9b-20241220"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "ae72fbc40f9c8afd52d1688b2233e5c915571de33618cfd1ffcb7f3fadfb4402", "supports": ["identity"], "url": "https://huggingface.co/api/models/THUDM/cogagent-9b-20241220"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "c1ad1d99a8d46a00610db35d97c787d10b7e81661e48a1e389ae6cbb51a45630", "supports": ["licenses.weights"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/LICENSE"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "de2c8e7a9792c3d129451ec813d458f83263460a5808cf87c45b04b260a376c9", "supports": ["licenses.weights"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:13Z", "role": "primary", "sha256": "35bbe72f95711e042dab7a94bf2711e30a3da50190c7315178c37de5725cc1d9", "supports": ["aliases"], "url": "https://huggingface.co/api/models/THUDM/cogagent-9b-20241220"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:14Z", "role": "primary", "sha256": "35bbe72f95711e042dab7a94bf2711e30a3da50190c7315178c37de5725cc1d9", "supports": ["aliases"], "url": "https://huggingface.co/api/models/zai-org/cogagent-9b-20241220"}], "id": "url:c5eb975c0d66f8a2954f", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/cogagent-9b-20241220"}], "name": "CogAgent 9B 20241220", "notes": "The former THUDM model URL resolves to zai-org. Six public safetensors shards are present. Weights use the custom CogAgent License, including a registration requirement for commercial use, rather than a blanket permissive license. The card cites the 2023 CogAgent paper as previous work and links a later technical report; no claim that the old paper documents this 2024 checkpoint is asserted.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md", "target_id": "github:thudm/cogagent", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "A bilingual GLM-4V-based execution model that uses screenshots, task instructions and action history to predict GUI operations.", "topics": ["grounding", "action-prediction", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/zai-org/cogagent-9b-20241220"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "computed", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "population": "sum of icon/text counts in the published card table", "qualifier": "exact", "unit": "target annotations", "value": 1581}], "evidence_urls": ["https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"], "formats": ["JSON annotations", "screen images"], "modalities": ["high-resolution screenshots", "instructions", "bounding boxes"], "notes": "Counts describe target annotations, not unique screenshots or expanded language/style variants. The evaluation code can reuse a target across configurations.", "splits": [], "supervision": ["GUI grounding"]}, "download_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/tree/main"}, "environments": ["desktop", "windows", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:38:05Z", "role": "primary", "sha256": "a77c859775b33591b5619e8520c94702a53a1f9a29ada93e66744dbe23afe9dc", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:10Z", "role": "primary", "sha256": "2f7f635dee3c593b40efc61246c082ab1d76f7ce78f84a1d4b6391d30fa4e8fd", "supports": ["relations"], "url": "https://gui-agent.github.io/grounding-leaderboard/"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}], "id": "url:c79e152e6f7af1a9efed", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:35:55Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-16T11:35:55Z", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro"}], "name": "ScreenSpot-Pro data", "notes": "This is the benchmark data asset, distinct from its evaluation-code repository. Target counts are not unique image counts.", "published_at": null, "relations": [{"evidence_url": "https://gui-agent.github.io/grounding-leaderboard/", "target_id": "arxiv:2504.07981", "type": "documented_by"}, {"evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "target_id": "github:likaixin2000/screenspot-pro-gui-grounding", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "High-resolution professional interface screenshots and target annotations for evaluating instruction-based GUI localization.", "topics": ["grounding", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/zonghanHZH/ZonUI-3B/tree/07212ea7724a69400422571e0f6e053212e66fe6", "gated": false, "hub_id": "zonghanHZH/ZonUI-3B", "hub_revision": "07212ea7724a69400422571e0f6e053212e66fe6", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"], "reason": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate."}}, "repository_url": "https://github.com/Han1018/ZonUI-3B"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "0dcd557508c07c9569e642a0319e71162b50c7ac6a0449564760ac2d6fffb089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:22Z", "role": "primary", "sha256": "c419027b878ed7ebebd91f89a7395506d2f2aa8e1745918833cf1bac9e0a054e", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/zonghanHZH/ZonUI-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "9918231748808ea1af93b559b0a511f1f9687689766a09b43c7726f14f27248b", "supports": ["release_status", "details.repository_url"], "url": "https://raw.githubusercontent.com/Han1018/ZonUI-3B/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:02Z", "role": "primary", "sha256": "7e3e63c1f2b38daaf4bb222b51fa5ae6594a0185e618723371e044bee6f05f7d", "supports": ["notes"], "url": "https://api.github.com/repos/Han1018/ZonUI-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:20Z", "role": "primary", "sha256": "d4c1500dbaceba738565d766e7b7453abbb6c0b12edaa34e5600d351214bfe4f", "supports": ["notes"], "url": "https://api.github.com/repos/Han1018/ZonUI-3B/git/trees/main?recursive=1"}], "id": "url:c8c046847f829fea4821", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:35:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:16Z", "url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zonghanHZH/ZonUI-3B"}], "name": "ZonUI-3B checkpoint", "notes": "The Hub manifest lists four-part PyTorch and Safetensors files as alternative representations of one checkpoint, plus configuration/tokenizer files. The card’s example rescales predicted locations through the resized image dimensions. No checkpoint was loaded and no benchmark score was reproduced. Apache-2.0 is the model-card weight declaration; the separate code repository has no project-level license in the inspected tree.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md", "target_id": "arxiv:2506.23491", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Qwen2.5-VL-3B-based GUI grounding checkpoint that predicts click locations from screenshots and text instructions across desktop, mobile and web screens.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": null, "url": "https://huggingface.co/zonghanHZH/ZonUI-3B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/tencent/POINTS-GUI-G/tree/main", "gated": false, "hub_id": "tencent/POINTS-GUI-G"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "ac679d3cfdda8754e945d127935c9ed9030e6c9d0c9872f6224909ab2abe5c3f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "3b97405c0a7f50298cf31a4933d31da356363c2ba5ecfaf73144cedeebc44db2", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/tencent/POINTS-GUI-G"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "9f851237932688c15d261c016d51ff5ae80c43c5d0b4eaccac0b48d941118d80", "supports": ["environments"], "url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "ac679d3cfdda8754e945d127935c9ed9030e6c9d0c9872f6224909ab2abe5c3f", "supports": ["licenses.weights"], "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}], "id": "url:cc931199bda2bdf1f5f7", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/tencent/POINTS-GUI-G"}], "name": "POINTS-GUI-G 8B", "notes": "Four safetensors shards and model code/configuration are public and ungated. The card declares license other without specifying a named weight license. This release is the grounding model; the same card says the end-to-end GUI agent model is still under development. No end-to-end agent capability is inferred from grounding scores.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md", "target_id": "arxiv:2602.06391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A specialized interface-grounding checkpoint that returns normalized point coordinates for a requested element in a screenshot.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/tencent/POINTS-GUI-G"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/microsoft/Phi-Ground/tree/main", "gated": false, "hub_id": "microsoft/Phi-Ground", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "ce8e0ebc3d4c2642bfb6a1d807176eae4c14b52fbc6f221b3c234a9d55e0d362", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/microsoft/Phi-Ground"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "ce8e0ebc3d4c2642bfb6a1d807176eae4c14b52fbc6f221b3c234a9d55e0d362", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/Phi-Ground"}], "id": "url:cee45b940d5fbff3ccca", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/microsoft/Phi-Ground"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/Phi-Ground"}], "name": "Phi-Ground checkpoint", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/microsoft/Phi-Ground", "target_id": "github:microsoft/phi-ground", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "Microsoft Phi-Ground GUI grounding checkpoint fine-tuned from Phi-3.5-vision-instruct, shipped as two PyTorch weight shards with custom Phi-3-V configuration and processing code under MIT.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/microsoft/Phi-Ground"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/cckevinn/SeeClick/tree/main", "gated": false, "hub_id": "cckevinn/SeeClick", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web", "desktop", "android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:52Z", "role": "primary", "sha256": "bc9f7c36f3e7c4515fa4982b0dd921a9038ff082cb6b283572d6390ef651c2a9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/cckevinn/SeeClick"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "c4112415bec58fc77f9a1e8fa98f8d602c5d382747579dfda9f866900c55995c", "supports": ["summary", "environments", "relations", "licenses"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:05:52Z", "role": "primary", "sha256": "bc9f7c36f3e7c4515fa4982b0dd921a9038ff082cb6b283572d6390ef651c2a9", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/cckevinn/SeeClick"}], "id": "url:cfea8a0949472a637990", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/cckevinn/SeeClick"}], "name": "SeeClick checkpoint", "notes": "The public ungated Hub manifest contains ten safetensors shards and model configuration/code but no README model card or explicit weight-license field. The official repository identifies the checkpoint and says incorporated checkpoints follow their original licenses, without resolving a specific standalone weight license. No source or model code was executed.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md", "target_id": "arxiv:2401.10935", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The released SeeClick visual grounding checkpoint for mapping an interface screenshot and referring instruction to a point or bounding box.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/cckevinn/SeeClick"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI/tree/main", "gated": false, "hub_id": "Miaosen/Grounding-Anything-GUI", "languages": ["en"], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "c64f72eae3dd50d28996f99ab744a5093896e8bb29caa6f7b33a2f60a0b530a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/Miaosen/Grounding-Anything-GUI"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "31b964cfd198f99491f0f480024b2322fbdf3a237f7806847e83978109bc5856", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a0609394a736dc495d077cd57a2945b018123e5b3f9514d7d5bafb9e690aeae2", "supports": ["summary", "release_status", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "c64f72eae3dd50d28996f99ab744a5093896e8bb29caa6f7b33a2f60a0b530a5", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Miaosen/Grounding-Anything-GUI"}], "id": "url:d1830de02226155ee16d", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI"}], "name": "Grounding-Anything-GUI", "notes": "The Hub exposes train.zip and a card containing only license/language metadata. The official repository marks this training-data release complete, but its Phi-Ground-Any paper is still pending. No binding to the earlier Phi-Ground paper, sample count, split structure or operating-system coverage is inferred.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md", "target_id": "github:microsoft/phi-ground", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "The public training archive linked by the Phi-Ground project for its Phi-Ground-Any model.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/tree/main", "gated": false, "hub_id": "InfiX-ai/InfiGUI-G1-7B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "253055f0f77266a64d2db5de8fd04c4202f4cda3be733dd1b00d8565116dd240", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "10278eb0f1600557e432a0f8dde2c40cfffe12330ddb949b6a0fc02bfbc6c9ee", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-G1-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "4b68c82a2dd0063a450bde34c8bffbd931565a91a2982a5b47be7e711815b284", "supports": ["environments"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "253055f0f77266a64d2db5de8fd04c4202f4cda3be733dd1b00d8565116dd240", "supports": ["licenses.weights"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}], "id": "url:d70a755e3b2102f25f34", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B"}], "name": "InfiGUI-G1 7B", "notes": "Four safetensors shards are public. The model card declares Apache-2.0 and identifies Qwen2.5-VL-7B-Instruct as its base. It is the G1 grounding checkpoint, distinct from the previously cataloged InfiGUI-R1 reasoning checkpoint; benchmark scores remain publisher claims.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md", "target_id": "arxiv:2508.05731", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A GUI-grounding checkpoint trained with adaptive exploration rewards and multiple candidate answers to improve instruction-to-element alignment.", "topics": ["grounding", "reinforcement-learning"], "updated_at": null, "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3-VL-8B-Instruct", "gated": false, "hub_revision": "9db378189fddaabb5f6e347b9053858a392022ac", "listed_weight_files": ["model.safetensors"]}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "73a4690fab49d3b08ea9b0b1e44d62d5ec1802614f5cf22c37ed92dca49aa520", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "8b680b9a63e654fe0b43754201d30c91714859578f294f4781e035280e862657", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/PrentisAI/GroundGUI-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "73a4690fab49d3b08ea9b0b1e44d62d5ec1802614f5cf22c37ed92dca49aa520", "supports": ["licenses.weights"], "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}], "id": "url:de31b07ff24ab15d9bdf", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/PrentisAI/GroundGUI-8B"}], "name": "GroundGUI-8B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "github:prentisai/groundgui", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "url:56bda04bd8d3a438deb3", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screenshot grounding checkpoint that emits a single click in a normalized coordinate space.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/PrentisAI/GroundGUI-8B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/ritzzai/GUI-R1/tree/main", "gated": false, "hub_id": "ritzzai/GUI-R1"}, "environments": ["desktop", "windows", "linux", "macos", "android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "f31ae9547d94967dd91f5cde5e5c66430d7d59ee3a525c0914bc600a79999af7", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/ritzzai/GUI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["summary", "environments", "release_status"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}], "id": "url:df680385f18b01a825be", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/ritzzai/GUI-R1"}], "name": "GUI-R1 data bundle", "notes": "The Hub is a dataset repository with Parquet payloads, although the short card mistakenly calls it a model. The official project identifies it as training/evaluation data and describes subsequent expanded mixtures. MIT is declared in this dataset card; upstream dataset terms remain separate. No 3K/800K aggregate count is inferred for the mutable bundle.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training and evaluation Parquet bundles accompanying GUI-R1, including GUI grounding and mobile/web action-prediction data.", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/ritzzai/GUI-R1"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xlangai/Jedi/raw/main/README.md"], "formats": [], "modalities": ["images", "text", "coordinates"], "notes": "Coordinates were resized for a specified Qwen2.5-VL image processor and must remain aligned with image resizing. AGUVIS++ material requires images from upstream AGUVIS and OS-Atlas releases; this repository is not a self-contained copy of every referenced image collection.", "splits": [], "supervision": ["icon grounding", "component grounding", "layout grounding", "refusal"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/Jedi/tree/main", "gated": false, "hub_id": "xlangai/Jedi", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T09:33:18Z", "role": "primary", "sha256": "63ff72bea0a2cb9218093d7a5a420b39c01861aef66141433d7c56c1e0ad05ad", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/xlangai/Jedi"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:38Z", "role": "primary", "sha256": "1d4d15c0cd27837602d55f0fda1ed91d9e8a98b33b3e3c4ba64decae42399558", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/Jedi/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T09:33:18Z", "role": "primary", "sha256": "63ff72bea0a2cb9218093d7a5a420b39c01861aef66141433d7c56c1e0ad05ad", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xlangai/Jedi"}], "id": "url:e062cf1445eaee0f3ade", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T09:33:18Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T09:33:18Z", "url": "https://huggingface.co/datasets/xlangai/Jedi"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/Jedi"}], "name": "Jedi", "notes": "The dataset card requires coordinate scaling to stay aligned with the image processor. Follow the publisher recipe before using the grounding labels.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/xlangai/Jedi", "target_id": "arxiv:2505.13227", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Synthetic GUI grounding data associated with UI decomposition research and the OSWorld-G benchmark.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/xlangai/Jedi"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 train", "qualifier": "exact", "unit": "question–answer examples", "value": 307315}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 train", "qualifier": "exact", "unit": "web pages", "value": 4549}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 dev", "qualifier": "exact", "unit": "question–answer examples", "value": 52826}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 dev", "qualifier": "exact", "unit": "web pages", "value": 913}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 test", "qualifier": "exact", "unit": "question–answer examples", "value": 40357}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 test", "qualifier": "exact", "unit": "web pages", "value": 985}], "evidence_urls": ["https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "https://huggingface.co/api/datasets/X-LANCE/WebSRC_v1.0", "https://x-lance.github.io/WebSRC/"], "formats": ["CSV", "HTML", "JSON", "ZIP packages"], "modalities": ["questions and answers", "HTML", "screenshots", "layout metadata"], "notes": "Each dataset.csv row is a QA example; web pages and QA examples are different units. Use the v1.0 split table rather than merging the older project headline into it. The card retains email-evaluation instructions, while the manifest now includes test answers; no submission was made. Payload archives were not downloaded.", "splits": ["train", "dev", "test"], "supervision": ["extractive or yes/no answers", "answer DOM-element references"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/tree/7aa0bc6efc7ef43f68c192e2091108541acbaf1a", "gated": false, "hub_id": "X-LANCE/WebSRC_v1.0", "hub_revision": "7aa0bc6efc7ef43f68c192e2091108541acbaf1a", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "061e5bcaa49e819f0133afb912ad5ce20a8c2b6c28ca60319a4635d6ce835ddb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:29Z", "role": "primary", "sha256": "a7976dd931c3fcf11de2456614f95eb071978599bd70fb89c0ea009d31da4300", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/X-LANCE/WebSRC_v1.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:04Z", "role": "primary", "sha256": "3c2c601d8b4ad2f379cac9b97a2cbb901883d9bc6c1a682359f9b4135e6dd4bc", "supports": ["notes", "details.dataset_profile"], "url": "https://x-lance.github.io/WebSRC/"}], "id": "url:e0ba328ae1a643b0a5aa", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:35:16Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-18T04:35:16Z", "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0"}], "name": "WebSRC v1.0", "notes": "Public release includes train+dev and test ZIP packages plus a test-answer JSON file. The v1.0 card’s exact split table differs from the older project-page headline of 0.44M questions/6.5K pages; counts are kept with their sources. This is static webpage understanding, not an interactive browser-agent trajectory dataset. CC-BY-4.0 is the data-card declaration.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "target_id": "arxiv:2101.09465", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Webpage segments, questions, answers, HTML, screenshots and layout metadata for structural reading comprehension.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/microsoft/OmniParser-v2.0/tree/main", "gated": false, "hub_id": "microsoft/OmniParser-v2.0", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect/LICENSE", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect_v3/LICENSE", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_caption/LICENSE"], "reason": "The top-level Hub metadata says MIT, but the same card and folder licenses distinguish the original detector from the new detector and captioner. Preserve one model-bundle identity with a mixed license declaration."}}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "f4ebaafcc0ab0b9e9cc55cb587678e5c83a5bbd64469fb2cc52bc63843c5cc6d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "1ea8af0cfd9cd716de8af8b78b2451470e1e1f22233a7286c5f40e1f496dc4be", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/OmniParser-v2.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "6ae154d65bfc2755703440be02acf63a37b741a1101b7c42b666da93a4ed2310", "supports": ["summary", "licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:09:13Z", "role": "primary", "sha256": "4df3c306dddaaf4baffdff5ca820cc679ac8cd6dc263c6a74517783e42fa7a3b", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "646f8936b8ddcd14e13e578ff6857e368780b0d1a4f6066bee89211923a373e2", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect_v3/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:09:13Z", "role": "primary", "sha256": "d9a1b1e30d633d5732ea18e3cba9538d293ebc53e1a9e4e96ab739e0c5c4f1cb", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_caption/LICENSE"}], "id": "url:e21ca7c859544dbf5fb6", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher per-folder model license declarations and retained LICENSE files; this is a component map, not one SPDX license for the full bundle.", "declared": "Mixed by component: icon_detect = AGPL-3.0; icon_detect_v3 = MIT; icon_caption = MIT", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/OmniParser-v2.0"}], "name": "OmniParser v2 model bundle", "notes": "The manifest contains icon_detect/model.pt, icon_detect_v3/model.pt and icon_caption/model.safetensors. Licenses differ by folder: original YOLOv8 detector AGPL-3.0; YOLOv9-E detector and Florence-2 captioner MIT. The card’s top-level MIT metadata must not cover the entire bundle. Components remain one resource; no binaries were executed.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A screen-parsing weight bundle containing icon-region detectors and an icon captioner for converting screenshots into agent-readable interface elements.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://huggingface.co/microsoft/OmniParser-v2.0"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL/tree/main", "gated": false, "hub_id": "microsoft/GUI-Actor-7B-Qwen2-VL", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "c76d9811dc6152d1b33bec7f19e808c1882e86cca79bd3fea90075ac9df5c7e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "c76d9811dc6152d1b33bec7f19e808c1882e86cca79bd3fea90075ac9df5c7e6", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL"}], "id": "url:e3e5dd82a3102689c4ea", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL"}], "name": "GUI-Actor 7B Qwen2-VL", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL", "target_id": "github:microsoft/gui-actor", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "GUI-Actor checkpoint fine-tuned from Qwen2-VL-7B-Instruct, with about 8.4B BF16 parameters, whose config names a pointer variant of the Qwen2-VL class rather than the stock one, under MIT.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/osunlp/UGround-V1-7B/tree/main", "gated": false, "hub_id": "osunlp/UGround-V1-7B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "080c9245276e9ffe5911551a66acd7e91b478f85b96c795d6079465d92580f18", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "080c9245276e9ffe5911551a66acd7e91b478f85b96c795d6079465d92580f18", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B"}], "id": "url:e554295a7f3f5c9ad0d6", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/osunlp/UGround-V1-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/osunlp/UGround-V1-7B"}], "name": "UGround V1 7B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B", "target_id": "github:osu-nlp-group/uground", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "UGround-V1 checkpoint fine-tuned from Qwen2-VL-7B, with about 8.3B BF16 parameters in the standard Qwen2-VL architecture, provided as safetensors and a PyTorch file under Apache-2.0.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/osunlp/UGround-V1-7B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "single-turn portion", "qualifier": "approximately", "unit": "QA pairs", "value": 44000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "multi-turn portion", "qualifier": "approximately", "unit": "dialogues", "value": 6000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "text", "bounding boxes"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": [], "supervision": ["single-turn question answering", "multi-turn dialogue"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIChat/tree/main", "gated": false, "hub_id": "yiye2023/GUIChat", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIChat/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "3ccb4a9ded4f45aa334f0df5dc9a0f8963c645857e8442c46efe1d9cf4a4360b", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIChat"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "4bcf87ecfbbb8e07a01b21415a970c8b53a5283bf6872b657040d3f45c9241f7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIChat/raw/main/README.md"}], "id": "url:edf38d43800be8c955ae", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIChat"}], "name": "GUIChat", "notes": "Public JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Conversational GUI supervision with text-rich images and bounding boxes, organized as single-turn questions and multi-turn dialogues.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/yiye2023/GUIChat"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/tree/main", "gated": false, "hub_id": "LZXzju/Qwen2.5-VL-3B-UI-R1"}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "620fbee20a6002bbc13562030ce4988d41ef17b0c033ec6872d83741f597adb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "6a206702a1a4faef428460e07f52aaf8fd1f1a144d711d807ca42479b248f697", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/LZXzju/Qwen2.5-VL-3B-UI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["environments"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "620fbee20a6002bbc13562030ce4988d41ef17b0c033ec6872d83741f597adb8", "supports": ["licenses.weights"], "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}], "id": "url:f13474d7aa734a30624d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1"}], "name": "UI-R1 3B", "notes": "Two safetensors shards and tokenizer/configuration files are publicly ungated. The model card declares MIT; upstream base-model terms are separate. The mutable card compares v1, v2 and UI-R1-E results without pinning every result to a commit, so this record identifies the Hub release and does not import an unversioned score.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL checkpoint trained with reinforcement learning for GUI action prediction and evaluated on mobile, desktop and web grounding tasks.", "topics": ["grounding", "reinforcement-learning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo1-7B/tree/main", "gated": false, "hub_id": "Hcompany/Holo1-7B"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "a7e185380bb02150d761651420cccc9c4a57f67c70a5dec3b7f85af63fb8773b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "66076674ebbdcf9faeb408ebd137ac1283fb56b1011acd7037ebb73e09ba749d", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo1-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "a7e185380bb02150d761651420cccc9c4a57f67c70a5dec3b7f85af63fb8773b", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}], "id": "url:f35cfcbaeb79663eabeb", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:50Z", "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo1-7B"}], "name": "Holo1 7B", "notes": "Four public safetensors shards and example navigation/localization files are present. The model card declares Apache-2.0. Reported Surfer-H results depend on the assembled policy/localizer/validator configuration and are not imported as standalone checkpoint scores.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md", "target_id": "arxiv:2506.02865", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based action model released for the Surfer-H web-agent system, supporting policy, interface-localization and validation roles.", "topics": ["web-navigation", "grounding", "planning"], "updated_at": null, "url": "https://huggingface.co/Hcompany/Holo1-7B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/showlab/ShowUI-pi/tree/main", "gated": false, "hub_id": "showlab/ShowUI-pi"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "64993ab5dc8205e25bda2576055b7e252104c67d4e6eeecb33a1a9742415f67e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/showlab/ShowUI-pi/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "ae12ece91946d79c5c6b578d6db07b5f095822be1ca3bc14bb3dc04a771bb010", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/showlab/ShowUI-pi"}], "id": "url:f70aa7c95824a8729516", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/showlab/ShowUI-pi"}], "name": "ShowUI-π checkpoint", "notes": "A model.safetensors file and policy pre/postprocessor configuration are publicly ungated. The card describes a SmolVLA-based GUI drag-and-drop model despite its generic robotics pipeline tag. No weight-license declaration or specific operating-system coverage is provided.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/showlab/ShowUI-pi/raw/main/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A compact visual-action checkpoint that predicts continuous GUI drag trajectories from a screenshot and a natural-language instruction.", "topics": ["grounding", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/showlab/ShowUI-pi"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "web subset", "qualifier": "approximately", "unit": "instructions", "value": 2900000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "web subset", "qualifier": "approximately", "unit": "images", "value": 173000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "mobile subset", "qualifier": "approximately", "unit": "instructions", "value": 1100000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "mobile subset", "qualifier": "approximately", "unit": "images", "value": 104000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "desktop subset", "qualifier": "approximately", "unit": "instructions", "value": 150000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "desktop subset", "qualifier": "approximately", "unit": "images", "value": 7800}], "evidence_urls": ["https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"], "formats": ["JSON annotations", "image archives"], "modalities": ["screenshots", "instructions", "element captions"], "notes": "The card identifies deprecated desktop files with screenshot-mapping errors and names corrected releases. Counts are rounded publisher figures for separate platform subsets; instructions and images are not interchangeable.", "splits": [], "supervision": ["element grounding", "element descriptions"]}, "download_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/tree/main", "modalities": ["image", "text"]}, "environments": ["web", "android", "desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "8713ca2a6b1962d5dce7da132e1135898d467053dc6b53b4463a0d43a6433bcb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "8713ca2a6b1962d5dce7da132e1135898d467053dc6b53b4463a0d43a6433bcb", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}], "id": "url:f8566bd0394d36ed4f8e", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T10:42:13Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-16T10:42:13Z", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data"}], "name": "Aria-UI Data", "notes": "The publisher flags original desktop files as deprecated because of screenshot-mapping issues, and identifies corrected versions in the dataset card.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI element descriptions and grounding instructions spanning web, mobile, and Ubuntu desktop interfaces.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/h-siyuan/ScreenDrag/raw/main/README.md"], "formats": ["Parquet", "episode files"], "modalities": ["screen images", "instructions", "previous state", "continuous actions"], "notes": "Raw and Parquet folders are alternative views of the same data. Each sample describes a dragging task over multiple timesteps; action vectors encode button state and coordinates.", "splits": ["train", "test"], "supervision": ["continuous GUI dragging"]}, "download_url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag/tree/main", "modalities": ["image", "text", "continuous-actions"]}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:12:56Z", "role": "primary", "sha256": "39c7c028f7ef5ca5c75f79b59a107c7e25444776095a2a08bb7535c4f523f7fa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "88df552ccbe80efd7020905d77d526f4f9efe2d09a11fed2714b083988f9dcef", "supports": ["relations"], "url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md"}], "id": "url:fe43af1801032be708cc", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:12:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag"}], "name": "ScreenDrag", "notes": "Raw and Parquet folders are alternative views of the same data and should not be counted as independent samples. The inspected card did not state a license.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md", "target_id": "github:showlab/showui-pi", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "GUI dragging data with screen observations, prior state, and continuous pointer actions, provided as trajectory files and Parquet representations.", "topics": ["grounding", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-2B", "gated": false, "hub_revision": "21e684b39c33e7efaf055f000374d74f37556954", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "08d14b9dc3010365e543cbb6707b67c5d792118451bbc2d0b70f1f0280cb20e1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "7b3242c3943537b6a26d8650a78c45044dce06a031ca2ce8bc67d5e482a332ce", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-2B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "08d14b9dc3010365e543cbb6707b67c5d792118451bbc2d0b70f1f0280cb20e1", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}], "id": "url:fec1c5aa8fd93cb78923", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B"}], "name": "Seg2Click Qwen3.5 2B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Seg2Click checkpoint using a 2B Qwen3.5 backbone to predict a mask and select an interface click.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/tree/main", "gated": false, "hub_id": "OpenGVLab/ScaleCUA-7B"}, "environments": ["web", "desktop", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e45116466b8b9ef3e93992a1efca8fa5b921a8b041297933210c9f552767ffb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:19Z", "role": "primary", "sha256": "bad2617fd64633212a30b78d99fb053a662378ab50b42722a04af4104ac2248f", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/OpenGVLab/ScaleCUA-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:54:55Z", "role": "primary", "sha256": "0538701c860afedd5d63c3839039699d8e5a14977d5c339b187ed13ad666c8fa", "supports": ["environments", "relations"], "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e45116466b8b9ef3e93992a1efca8fa5b921a8b041297933210c9f552767ffb8", "supports": ["licenses.weights"], "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}], "id": "url:ff7b1873a3eead89607a", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:45:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B"}], "name": "ScaleCUA 7B", "notes": "Four public safetensors shards and model/tokenizer configuration are present. The card declares Apache-2.0. The model card’s Paper link points to the code repository, so the arXiv association uses the repository’s explicit citation rather than treating that link as a paper URL.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md", "target_id": "arxiv:2509.15221", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based computer-use checkpoint with direct grounding and agent action modes across desktop, mobile and web interfaces.", "topics": ["grounding", "action-prediction", "planning"], "updated_at": null, "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B"}