仅在必要时辩论：面向高效大语言模型推理的自适应多智能体协作
Debate Only When Necessary: Adaptive Multiagent Collaboration for Efficient LLM Reasoning
AI总结：
针对多智能体协作提升LLM推理时存在的计算开销大、不必要辩论易引发错误的问题，提出自适应框架DOWN，按初始响应置信度选择性启动辩论，在基准测试中效率最高提升6倍且性能更优，降低了错误传播风险。
摘要：
多智能体协作已成为提升大语言模型（LLM）推理能力的一种极具前景的框架。尽管推理能力有所提升，但该方法因智能体间的迭代交互带来了巨大的计算开销。此外，参与不必要的辩论会增加生成错误响应的风险。为应对这些挑战，我们提出了“仅在必要时辩论”（Debate Only When Necessary, DOWN），这是一种自适应多智能体辩论框架，可根据智能体初始响应的置信度得分选择性地启动辩论。辩论仅针对需要进一步斟酌的查询启动，在此过程中，智能体通过参考同行的响应及相关置信度得分来优化自身输出。在多个基准上的评估表明，DOWN将效率提升了至多6倍，同时保持甚至超越了现有方法的性能。进一步分析显示，DOWN有效降低了由不必要的辩论过程引发的错误传播风险。这些发现证明了我们的方法在以更低计算成本提供高性能LLM解决方案方面的有效性。
英文摘要：
Multiagent collaboration has emerged as a promising framework for enhancing the reasoning capabilities of large language models (LLMs). Despite improvements in reasoning, the approach introduces substantial computational overhead resulting from iterative agent interactions. Furthermore, engaging in unnecessary debates increases the risk of generating erroneous responses. To address these challenges, we propose Debate Only When Necessary (DOWN), an adaptive multiagent debate framework that selectively activates debate based on the confidence score of the agent's initial response. Debate is activated only for queries requiring further deliberation, during which agents refine their outputs by referencing peer responses and associated confidence scores. Evaluations on benchmarks show that DOWN improves efficiency by up to six times while preserving or even outperforming the performance of existing methods. Further analysis indicates that DOWN effectively mitigates the risk of error propagation stemming from the unnecessary debate process. These findings demonstrate the effectiveness of our approach in delivering high-performance LLM solutions at a lower computational cost.
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。
