Metraipolis › archive › dc_28618b978dd3

LoongReflect: Boosting Long-Horizon Reflection in Search Agents via Global Perspective Distillation — arXiv:2608.11967

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.11967
License (as archived)
CC0-1.0
Archived
2026-10-19T05:16:19.662Z
SHA-256 of the archived text
622137bbf4d0aec6e436fff6f397f084b61a1688b54d6815e50a0f65e7356c50
Stable URL of this page
https://metraipolis.com/documento/dc_28618b978dd3

How to cite

LoongReflect: Boosting Long-Horizon Reflection in Search Agents via Global Perspective Distillation — arXiv:2608.11967 — arXiv (en), https://arxiv.org/abs/2608.11967 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_28618b978dd3 (Metraipolis archive dc_28618b978dd3, sha256 622137bbf4d0aec6e436fff6f397f084b61a1688b54d6815e50a0f65e7356c50)
BibTeX
@misc{dc28618b978dd3,
  title = {LoongReflect: Boosting Long-Horizon Reflection in Search Agents via Global Perspective Distillation — arXiv:2608.11967},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.11967},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_28618b978dd3; sha256 622137bbf4d0aec6e436fff6f397f084b61a1688b54d6815e50a0f65e7356c50},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.11967. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language model agents increasingly rely on long-horizon reasoning to solve complex tasks involving planning, tool use, and memory. A critical capability in such settings is reflection: assessing trajectory progress, identifying missing evidence and unreliable intermediate states, and deciding whether to continue, revise, or abandon the current branch. Learning effective reflection, however, is challenging because reflection is performed locally within the current branch, whereas its utility can only be determined by its contribution to the final trajectory outcome. This local-global mismatch makes outcome-based reinforcement learning provide only local, sparse and delayed supervision for reflective decisions. To solve these, we propose LoongReflect, a training framework that formulates reflection as a memory-control policy. The agent operates over a reversible trajectory tree using explicit reflect and backtrack actions. Reflection consolidates verified facts, missing evidence, and branch-specific risks into working memory, while backtracking removes an unreliable branch from the active context and preserves a concise corrective lesson. To learn this policy, LoongReflect combines two complementary signals through a look-ahead, extragradient-style coordination mechanism. A fast channel distills globally informed reflective behavior from a privileged teacher, with supervision restricted to reflection and backtracking tokens. A slow channel optimizes complete trajectories using outcome-based GRPO, aligning local control decisions with final task success. Experiments on multi-hop retrieval-augmented generation and mathematical reasoning benchmarks demonstrate consistent improvements over outcome-only reinforcement learning and self-distillation baselines.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-13 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.11967 — 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 2026-10-19T05:08:46.748Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-13?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).