arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2503.06951cs.AI
收藏

ReAgent:用于知识增强多跳问答的可逆多智能体推理

ReAgent: Reversible Multi-Agent Reasoning for Knowledge-Enhanced Multi-Hop QA

The University of Tokyo(东京大学);Duke–NUS Medical School(杜克-国立新加坡大学医学院);The Hong Kong University of Science and Technology(香港科学与技术大学)

Xinjie Zhao, Fan Gao, Xingyu Song, Yingjian Chen, Rui Yang, Yanran Fu, Yuyang Wang, Yusuke Iwasawa, Yutaka Matsuo, Irene Li

AI总结:

针对思维链推理错误不可逆累积的问题,提出可逆多智能体协作框架ReAgent,通过显式回溯、文本检索与信息验证实现推理中途纠错,在三个基准上平均提升约6%,为容错问答系统奠定基础。

摘要:

大型语言模型(LLMs)的最新进展通过直接的思维链(CoT)推理显著改善了多跳问答(QA)。然而,CoT的不可逆性导致错误累积,使得纠正多跳推理中的错误极具挑战。本文提出ReAgent:一个增强显式回溯机制的可逆多智能体协作框架,实现可逆的多跳推理。通过结合基于文本的检索、信息聚合与验证,该系统能在推理中途检测并纠正错误,从而产生更稳健且可解释的QA结果。该框架与实验为未来容错QA系统的研究奠定基础。在三个基准上的经验评估表明了ReAgent的有效性,相较基线模型平均提升约6%。

英文摘要:

Recent advances in large language models (LLMs) have significantly improved multi-hop question answering (QA) through direct Chain-of-Thought (CoT) reasoning. However, the irreversible nature of CoT leads to error accumulation, making it challenging to correct mistakes in multi-hop reasoning. This paper introduces ReAgent: a Reversible multi-Agent collaborative framework augmented with explicit backtracking mechanisms, enabling reversible multi-hop reasoning. By incorporating text-based retrieval, information aggregation and validation, our system can detect and correct errors mid-reasoning, leading to more robust and interpretable QA outcomes. The framework and experiments serve as a foundation for future work on error-tolerant QA systems. Empirical evaluations across three benchmarks indicate ReAgent's efficacy, yielding average about 6\% improvements against baseline models.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑