arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.11527cs.AI
收藏

改进LLM推理的多智能体思维树验证器智能体

Improving LLM Reasoning with Multi-Agent Tree-of-Thought Validator Agent

University of Texas at San Antonio(德克萨斯大学圣安东尼奥分校);Peraton Labs(佩拉顿实验室)

Fatemeh Haji, Mazal Bethany, Maryam Tabar, Jason Chiang, Anthony Rios, Peyman Najafirad

AI总结:

提出结合多智能体思维树推理与思维验证器的方法,通过丢弃无效推理路径实现稳健投票,在GSM8K上平均提升5.6%。

摘要:

多智能体策略通过为问题解决过程分配专门角色,已成为增强大型语言模型(LLM)推理能力的一种有前景的方法。与此同时,思维树(ToT)方法通过探索多样化的推理路径,在改进复杂问答任务推理方面展现出潜力。多智能体推理中的一个关键限制是“推理者”智能体对推理路径的浅层探索。虽然ToT策略有助于缓解这一问题,但它们可能生成有缺陷的推理分支,这可能损害最终答案的可信度。为了利用多智能体推理和ToT策略两者的优势,我们引入了一种新颖方法,将基于ToT的推理者智能体与思维验证器智能体相结合。多个推理者智能体并行运行,采用ToT探索多样化的推理路径。随后,思维验证器审查这些路径,仅当推理者的推理有效时才考虑其结论。该方法通过丢弃有缺陷的推理路径,实现更稳健的投票策略,增强系统处理需要系统性和可信推理任务的能力。在GSM8K数据集上的评估表明,与现有技术相比,我们的方法表现出更优性能,在四个LLM上平均比标准ToT策略高出5.6%。相关代码和内容可在https://github.com/SecureAIAutonomyLab/MA-ToT找到。

英文摘要:

Multi-agent strategies have emerged as a promising approach to enhance the reasoning abilities of Large Language Models (LLMs) by assigning specialized roles in the problem-solving process. Concurrently, Tree of Thoughts (ToT) methods have shown potential in improving reasoning for complex question-answering tasks by exploring diverse reasoning paths. A critical limitation in multi-agent reasoning is the 'Reasoner' agent's shallow exploration of reasoning paths. While ToT strategies could help mitigate this problem, they may generate flawed reasoning branches, which could harm the trustworthiness of the final answer. To leverage the strengths of both multi-agent reasoning and ToT strategies, we introduce a novel approach combining ToT-based Reasoner agents with a Thought Validator agent. Multiple Reasoner agents operate in parallel, employing ToT to explore diverse reasoning paths. The Thought Validator then scrutinizes these paths, considering a Reasoner's conclusion only if its reasoning is valid. This method enables a more robust voting strategy by discarding faulty reasoning paths, enhancing the system's ability to tackle tasks requiring systematic and trustworthy reasoning. Our method demonstrates superior performance compared to existing techniques when evaluated on the GSM8K dataset, outperforming the standard ToT strategy by an average 5.6% across four LLMs. The code and related content can be found in: https://github.com/SecureAIAutonomyLab/MA-ToT

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。

↑