Metraipolis › archive › dc_a647cbe5fdc7

AgenticRag-R1: Agentic Reinforcement Learning with Stack Memory for Multi-Step Reasoning, Retrieval and Memorizing — arXiv:2608.29622

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.29622
License (as archived)
CC0-1.0
Archived
2026-10-17T07:40:49.947Z
SHA-256 of the archived text
e5dc21354d5254c4738dd6209fb7c5ef6bcbe32ca1537e93af06856131193a64
Stable URL of this page
https://metraipolis.com/documento/dc_a647cbe5fdc7

How to cite

AgenticRag-R1: Agentic Reinforcement Learning with Stack Memory for Multi-Step Reasoning, Retrieval and Memorizing — arXiv:2608.29622 — arXiv (en), https://arxiv.org/abs/2608.29622 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a647cbe5fdc7 (Metraipolis archive dc_a647cbe5fdc7, sha256 e5dc21354d5254c4738dd6209fb7c5ef6bcbe32ca1537e93af06856131193a64)
BibTeX
@misc{dca647cbe5fdc7,
  title = {AgenticRag-R1: Agentic Reinforcement Learning with Stack Memory for Multi-Step Reasoning, Retrieval and Memorizing — arXiv:2608.29622},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.29622},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a647cbe5fdc7; sha256 e5dc21354d5254c4738dd6209fb7c5ef6bcbe32ca1537e93af06856131193a64},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.29622. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Retrieval-Augmented Generation (RAG) improves the factuality of large language models (LLMs), yet existing RAG systems often struggle with complex, multi-step reasoning that requires adaptive retrieval and continuous revision of intermediate contexts. Recent reinforcement learning (RL)-based agentic RAG methods partially alleviate this issue, but typically rely on coarse-grained action spaces and trajectory-level rewards, resulting in weak reward assignment and a bias toward short-horizon, stereotyped reasoning template. To address, we propose AgenticRag-R1, a RL framework that deeply integrates reasoning, retrieval, and memory via a memory stack and fine-grained action space, supported by hierarchical action-aware rewards and an information-aware trajectory rejection strategy to enable effective long-horizon learning. Experiments across a diverse set of multi-hop, open-domain, and agentic reasoning benchmarks, spanning multiple backbone model sizes, demonstrate that AgenticRag-R1 consistently outperforms strong baselines. Moreover, AgenticRag-R1 learns more robust, interpretable, and memory-aware reasoning behaviors, highlighting the effect of fine-grained action modeling and information-aware optimization for long-horizon reasoning. Our code is anonymous available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.MA).
Categorie arXiv: Multiagent Systems (cs.MA) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.29622 — 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 2026-10-17T07:27:41.979Z dalla pagina https://arxiv.org/catchup/cs.MA/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).