Hear Both Sides: Efficient Multi-Agent Debate via Diversity-Aware Message Retention

2026-03-21

AI Summary

This work addresses the inefficiencies of full broadcast communication in multi-agent debate, which often introduces noise and redundancy, degrading reasoning quality and wasting computational resources. To mitigate these issues, the authors propose the Diversity-Aware Retention (DAR) framework, which selectively broadcasts messages from a subset of agents that exhibit maximal disagreement with both the group consensus and other participants at each debate round. Crucially, DAR explicitly preserves original divergent information through an index-based message retention mechanism, circumventing reliance on unreliable confidence thresholds. This lightweight and efficient communication strategy significantly enhances debate performance across multiple reasoning and question-answering benchmarks, demonstrating particular robustness in mitigating noise accumulation as the number of agents scales up.

Technology Category

• Multiagent Systems: Agent Communication
• Knowledge Representation and Reasoning: Argumentation
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Multi-Agent Debate has emerged as a promising framework for improving the reasoning quality of large language models through iterative inter-agent communication. However, broadcasting all agent messages at every round introduces noise and redundancy that can degrade debate quality and waste computational resources. Current approaches rely on uncertainty estimation to filter low-confidence responses before broadcasting, but this approach is unreliable due to miscalibrated confidence scores and sensitivity to threshold selection. To address this, we propose Diversity-Aware Retention (DAR), a lightweight debate framework that, at each debate round, selects the subset of agent responses that maximally disagree with each other and with the majority vote before broadcasting. Through an explicit index-based retention mechanism, DAR preserves the original messages without modification, ensuring that retained disagreements remain authentic. Experiments on diverse reasoning and question answering benchmarks demonstrate that our selective message propagation consistently improves debate performance, particularly as the number of agents scales, where noise accumulation is most severe. Our results highlight that what agents hear is as important as what agents say in multi-agent reasoning systems.

Problem

• Multi-Agent Debate
• Message Redundancy
• Noise Accumulation
• Confidence Calibration
• Selective Broadcasting

Innovation

• Multi-Agent Debate
• Diversity-Aware Retention
• Selective Message Propagation
• Disagreement Maximization
• Index-based Retention