Metraipolis › archive › dc_be011ede5285

DeReAct: Decomposed Reasoning and Acting for Reliable AI Agents — arXiv:2610.02351

Source
arXiv
Original
https://arxiv.org/abs/2610.02351
License (as archived)
CC0-1.0
Archived
2026-10-05T22:48:55.247Z
SHA-256 of the archived text
347c0469b89bb291802af596452770df3dbbf54a2dedbd4c772e678ab2c5a34e
Stable URL of this page
https://metraipolis.com/documento/dc_be011ede5285

How to cite

DeReAct: Decomposed Reasoning and Acting for Reliable AI Agents — arXiv:2610.02351 — arXiv, https://arxiv.org/abs/2610.02351 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_be011ede5285 (Metraipolis archive dc_be011ede5285, sha256 347c0469b89bb291802af596452770df3dbbf54a2dedbd4c772e678ab2c5a34e)
BibTeX
@misc{dcbe011ede5285,
  title = {DeReAct: Decomposed Reasoning and Acting for Reliable AI Agents — arXiv:2610.02351},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.02351},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_be011ede5285; sha256 347c0469b89bb291802af596452770df3dbbf54a2dedbd4c772e678ab2c5a34e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.02351. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

ReAct-based agents typically rely on a single LLM policy to propose actions, interact with the environment, and decide when a task is complete. This coupling makes action authorization and completion control difficult to enforce independently, allowing errors to propagate and unsupported completion claims to terminate execution. We introduce DeReAct, a modular agent architecture that externalizes two gating policies: a Critic that validates proposed actions before execution, and a Context Manager that reconstructs an environment-supported \textsc{State} and certifies task completion. Across GAIA and SWE-bench Verified, DeReAct improves Pass@1 most for weaker Brain models, with gains of 6.5--7.0 points for Qwen3-Coder-480B and 4.2--5.2 points for Claude Sonnet~4.5; gains diminish as Brain capability increases. Trajectory and ablation analyses show that external gating is effective when targeted failures are sufficiently prevalent and the gating policy is itself sufficient. With Claude Opus~4.5, Pass@1 remains comparable to ReAct, while DeReAct produces more evidence-complete and constraint-satisfying trajectories, indicating that completion control can trade earlier termination for stronger grounding. Overall, DeReAct improves weaker agents while retaining grounding benefits as models strengthen.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.02351 — 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 2026-10-05T22:45:55.577Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).