arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.12374cs.AIcs.SI
收藏

语言模型网络中的问题求解

Problem-Solving in Language Model Networks

University of Tsukuba(筑波大学);Grenoble INP - Ensimag, UGA(格勒诺布尔国立高等信息与应用数学学院)

Ciaran Regan, Alexandre Gournail, Mizuki Oka

AI总结:

本研究将多智能体辩论拓展至通用网络拓扑,发现随机网络性能与全连接网络相当且token消耗更少,揭示了自我反思与互联性的平衡及枢纽节点偏差的影响,为提升多智能体问答性能提供了新思路。

摘要:

为提升大语言模型(LLM)的推理与问答能力,学界已提出多种多智能体方法。尽管这些方法能提升性能,但基于集体智能的方法在复杂网络结构中的应用及智能体交互的动力学机制仍未得到充分探索。本研究将多智能体辩论的概念拓展至更通用的网络拓扑结构,衡量了集体的问答准确率、影响力、共识程度以及偏差带来的影响。结果显示,随机网络的表现与全连接网络相当,但其使用的token数量显著更少。此外,智能体间的强共识与正确答案相关,而分歧的回答通常意味着答案错误。对智能体影响力的分析揭示了自我反思与互联性之间的平衡:当局部交互给出错误答案时,自我反思能发挥作用;当智能体自身答案错误时,局部交互能提供帮助。另外,偏差对系统性能有显著影响,带有正确偏差的枢纽节点能提升性能。这些发现表明,使用随机网络或无标度网络并将知识丰富的智能体置于中心位置,可提升多智能体系统的整体问答性能。

英文摘要:

To improve the reasoning and question-answering capabilities of Large Language Models (LLMs), several multi-agent approaches have been introduced. While these methods enhance performance, the application of collective intelligence-based approaches to complex network structures and the dynamics of agent interactions remain underexplored. This work extends the concept of multi-agent debate to more general network topologies, measuring the question-answering accuracy, influence, consensus, and the effects of bias on the collective. The results show that random networks perform similarly to fully connected networks despite using significantly fewer tokens. Furthermore, a strong consensus among agents correlates with correct answers, whereas divided responses typically indicate incorrect answers. Analysing the influence of the agents reveals a balance between self-reflection and interconnectedness; self-reflection aids when local interactions are incorrect, and local interactions aid when the agent itself is incorrect. Additionally, bias plays a strong role in system performance with correctly biased hub nodes boosting performance. These insights suggest that using random networks or scale-free networks with knowledgeable agents placed in central positions can enhance the overall question-answering performance of multi-agent systems.

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。

↑