Metraipolis › archive › dc_135d2233464e
KnowGPO: Knowledge-Grouped Policy Optimization for Search Agents — arXiv:2610.22331
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.22331
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T09:05:57.004Z
- SHA-256 of the archived text
2667287c17f6ccd60a6ce94ce27aedda6d8c759d62eb6617c245e1f3ecf11c6a- Stable URL of this page
https://metraipolis.com/documento/dc_135d2233464e
How to cite
KnowGPO: Knowledge-Grouped Policy Optimization for Search Agents — arXiv:2610.22331 — arXiv (en), https://arxiv.org/abs/2610.22331 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_135d2233464e (Metraipolis archive dc_135d2233464e, sha256 2667287c17f6ccd60a6ce94ce27aedda6d8c759d62eb6617c245e1f3ecf11c6a)
BibTeX
@misc{dc135d2233464e,
title = {KnowGPO: Knowledge-Grouped Policy Optimization for Search Agents — arXiv:2610.22331},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.22331},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_135d2233464e; sha256 2667287c17f6ccd60a6ce94ce27aedda6d8c759d62eb6617c245e1f3ecf11c6a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.22331. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Deep search agents solve knowledge-intensive tasks through long-horizon, multi-turn interactions, where assigning credit to individual search decisions is fundamentally difficult. Outcome-based reinforcement learning provides only a single terminal reward shared across an entire trajectory. Recent process-supervision methods aim to assign credit at finer granularity, but they typically anchor credit to either surface-level trajectory states or model-internal signals, both of which become unreliable in deep search where observations are heterogeneous and retrieved contexts are long and noisy. We propose KnowGPO, a reinforcement learning framework that aligns trajectories by the task-relevant knowledge they have acquired rather than by their surface observations. For each query, KnowGPO decomposes the information required to derive the answer into atomic knowledge conditions, and represents the state of a trajectory at each step by the subset of conditions it has satisfied. Trajectories sharing the same knowledge state are grouped for within-group comparison, and the incremental knowledge gain between consecutive steps serves as an environment-grounded process reward. Experiments on Xbench, BrowseComp-ZH, and GAIA show KnowGPO consistently outperforms outcome-only (GRPO) and process-supervision baselines (GiGPO, IGPO) by 3.8 to 4.2 points, with our 8B model rivaling or exceeding most 32B open-source systems. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.22331 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T08:44:17.840Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).