Saved in:
Bibliographic Details
Main Authors: Li, Yunxuan, Du, Yibing, Zhang, Jiageng, Hou, Le, Grabowski, Peter, Li, Yeqing, Ie, Eugene
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2406.11776
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910491330215936
author Li, Yunxuan
Du, Yibing
Zhang, Jiageng
Hou, Le
Grabowski, Peter
Li, Yeqing
Ie, Eugene
author_facet Li, Yunxuan
Du, Yibing
Zhang, Jiageng
Hou, Le
Grabowski, Peter
Li, Yeqing
Ie, Eugene
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2406_11776
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Improving Multi-Agent Debate with Sparse Communication Topology
Li, Yunxuan
Du, Yibing
Zhang, Jiageng
Hou, Le
Grabowski, Peter
Li, Yeqing
Ie, Eugene
Computation and Language
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.
title Improving Multi-Agent Debate with Sparse Communication Topology
topic Computation and Language
url https://arxiv.org/abs/2406.11776