# AMBER: Training Long-Horizon Web Agents through Append-Only Memory

Introduces AMBER, an append-only memory framework in which a web agent learns to reason, act, and retain evidence using outcome-reward reinforcement learning. It studies whether preserving entries avoids the loss of facts and corrective feedback seen with rewritten memory.

- Catalog URL: https://awesomecua.github.io/resources/arxiv-2610-07118/
- Original source: https://arxiv.org/abs/2610.07118
- Kind: paper
- Review: curated
- Scope: core
- Environments: web
- Topics: memory, long-horizon, reinforcement-learning
- Evidence checked: 2026-10-16T06:36:29Z
- ID: arxiv:2610.07118

## Limitations

The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.

- Record JSON: https://awesomecua.github.io/resources/arxiv-2610-07118/record.json

## Authors

Chinmay Savadikar, Zhaoyu Zhang, Mingyu Zhao, Shuang Xie, Han Li, Tianfu Wu, Lingyun Wang

## Citation exports

- BibTeX: https://awesomecua.github.io/resources/arxiv-2610-07118/citation.bib
- CSL JSON: https://awesomecua.github.io/resources/arxiv-2610-07118/citation.csl.json

## Recorded arXiv snapshot

- Version: 1
- Version source: https://arxiv.org/abs/2610.07118v1
- First submission: 2026-10-05T17:11:07Z
- Version updated: 2026-10-05T17:11:07Z
- Abstract SHA-256: 4890856d5961395a9ac8c83c60b988da4adeaab1d99edd511e4046e4f99fd25d
- Abstract hash normalization: sha256-utf8-whitespace-collapsed-v1

## License declarations

- code: not recorded
- data: not recorded
- document: not recorded
- weights: not recorded

## Primary evidence

- https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%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 (retrieved 2026-10-16T06:36:29Z; SHA-256 74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f)
- https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%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 (retrieved 2026-10-16T06:36:29Z; SHA-256 74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f)