The Value of Variance: Mitigating Debate Collapse in Multi-Agent Systems via Uncertainty-Driven Policy Optimization

2026-02-06

AI Summary

This work addresses the problem of “debate collapse” in multi-agent debate systems, where erroneous reasoning often dominates due to the absence of effective detection and intervention mechanisms. The authors propose a hierarchical uncertainty quantification framework that measures behavioral uncertainty at the individual, interaction, and system levels, introducing it for the first time as a diagnostic indicator of system failure. Building on this, they develop an uncertainty-driven strategy optimization method that dynamically penalizes self-contradictory statements, inter-agent conflicts, and low-confidence outputs. Experimental results demonstrate that the proposed approach significantly improves decision accuracy, reduces internal inconsistency, and achieves reliable calibration of multi-agent systems across multiple benchmarks.

Technology Category

• Multiagent Systems: Multiagent Systems under Uncertainty
• Reasoning under Uncertainty: Uncertainty Representations
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Multi-agent debate (MAD) systems improve LLM reasoning through iterative deliberation, but remain vulnerable to debate collapse, a failure type where final agent decisions are compromised on erroneous reasoning. Existing methods lack principled mechanisms to detect or prevent such failures. To address this gap, we first propose a hierarchical metric that quantifies behavioral uncertainty at three levels: intra-agent (individual reasoning uncertainty), inter-agent (interactive uncertainty), and system-level (output uncertainty). Empirical analysis across several benchmarks reveals that our proposed uncertainty quantification reliably indicates system failures, which demonstrates the validity of using them as diagnostic metrics to indicate the system failure. Subsequently, we propose a mitigation strategy by formulating an uncertainty-driven policy optimization to penalize self-contradiction, peer conflict, and low-confidence outputs in a dynamic debating environment. Experiments demonstrate that our proposed uncertainty-driven mitigation reliably calibrates the multi-agent system by consistently improving decision accuracy while reducing system disagreement.

Problem

• debate collapse
• multi-agent systems
• uncertainty quantification
• LLM reasoning
• system failure

Innovation

• uncertainty quantification
• multi-agent debate
• debate collapse
• policy optimization
• behavioral uncertainty