利用稀疏通信拓扑改进多智能体辩论
Improving Multi-Agent Debate with Sparse Communication Topology
AI总结：
本文研究多智能体辩论中的全连接通信瓶颈，提出使用稀疏通信拓扑，在 GPT 与 Mistral 上以更低计算成本取得相当或更优表现，并扩展至多模态推理和对齐标注任务。
摘要：
多智能体辩论已被证明能有效提升大型语言模型在推理和事实性任务上的质量。尽管多智能体辩论中的各种角色扮演策略已得到探索，但在智能体之间的通信方面，现有方法采用的是一种暴力算法——每个智能体都能与所有其他智能体通信。本文系统研究了多智能体系统中通信连通性的影响。我们在 GPT 和 Mistral 模型上的实验表明，利用稀疏通信拓扑的多智能体辩论可以取得相当甚至更优的性能，同时显著降低计算成本。此外，我们将多智能体辩论框架扩展到多模态推理和对齐标注任务，展示了其广泛的适用性和有效性。我们的发现强调了通信连通性对于提升“心智社会”方法的效率和有效性的重要性。
英文摘要：
Multi-agent debate has proven effective in improving large language models quality for reasoning and factuality tasks. While various role-playing strategies in multi-agent debates have been explored, in terms of the communication among agents, existing approaches adopt a brute force algorithm -- each agent can communicate with all other agents. In this paper, we systematically investigate the effect of communication connectivity in multi-agent systems. Our experiments on GPT and Mistral models reveal that multi-agent debates leveraging sparse communication topology can achieve comparable or superior performance while significantly reducing computational costs. Furthermore, we extend the multi-agent debate framework to multimodal reasoning and alignment labeling tasks, showcasing its broad applicability and effectiveness. Our findings underscore the importance of communication connectivity on enhancing the efficiency and effectiveness of the "society of minds" approach.
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。
