Certifying Collective Reasoning in Multi-Agent Systems via Koopman Spectral Analysis
作者: Nuzhat Khan, Indrakshi Dey
分类: cs.MA, eess.SY
发布日期: 2026-08-06
💡 一句话要点
通过Koopman谱分析认证多智能体系统的集体推理
🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 多智能体系统 Koopman算子 集体推理 谱分析 智能决策 透明性
📋 核心要点
- 现有的多智能体系统在集体推理中缺乏透明性,无法有效测试收敛性和决策依据。
- 论文提出了一种基于Koopman算子理论的框架,通过谱分析来理解和验证集体推理的动态行为。
- 实验结果显示,该方法在96%的配置中准确预测了收敛时间,并在决策保真度上达到了99.7%。
📝 摘要(中文)
随着大型语言模型(LLM)智能体的集体辩论与投票形式的兴起,智能行为体现在相互作用中,而非单一智能体。尽管这种方法提高了任务准确性,但在系统层面仍然是黑箱,缺乏收敛性测试、所需轮次的界限以及决策驱动因素的清晰说明。本文基于Koopman算子理论开发了一种新框架,并验证了其在多智能体共识动态中的理论保证。将集体视为一个在通信图上的非线性动力系统,通过其Koopman转移算子的谱读取其基本行为,提供了三种可机器检查的证书,确保了集体推理的可信性。
🔬 方法详解
问题定义:本文旨在解决多智能体系统在集体推理中的黑箱问题,现有方法缺乏有效的收敛性测试和决策依据的透明性。
核心思路:通过引入Koopman算子理论,将集体视为一个非线性动力系统,从其谱中提取关键特征,以实现对推理过程的认证和分析。
技术框架:整体架构包括三个主要模块:1) 通过Koopman转移算子提取谱特征;2) 计算子主特征值以确定推理的内在时间尺度;3) 生成可机器检查的证书以验证推理的有效性。
关键创新:最重要的创新在于利用Koopman算子提供的谱特征,能够在推理前预测收敛时间,并通过谱证书验证推理的有效性,这在现有方法中是前所未有的。
关键设计:在实验中,使用了注意力共识模型,设置了特定的参数以确保谱分析的准确性,并通过15次辩论学习证书,确保在60个持出辩论中保持高保真度。
🖼️ 关键图片
📊 实验亮点
实验结果显示,所提出的方法在96%的配置中成功预测了收敛时间,并与观察到的收敛情况具有0.93的对数-对数相关性。此外,决策的保真度达到了99.7%,展现了该方法在实际应用中的有效性和可靠性。
🎯 应用场景
该研究的潜在应用领域包括智能决策系统、自动化辩论平台和多智能体协作任务等。通过提供可验证的推理过程,该方法能够增强系统的可信性和透明度,促进智能体在复杂环境中的协作与决策能力。
📄 摘要(原文)
Orchestrated collectives of large language model (LLM) agents that debate and vote are an emerging form of computational intelligence: the intelligent behaviour resides in the \emph{interaction}, not in any single agent. They improve task accuracy, yet remain black boxes at the system level: there is no principled test of convergence, no bound on the rounds needed, and no faithful account of what drove a decision. This paper develops a novel framework based on Koopman operator theory and validates its theoretical guarantees on multi-agent consensus dynamics. Treating the collective as one nonlinear dynamical system on a communication graph, we read its essential behaviour off the spectrum of its Koopman transfer operator, an exact linear representation of the nonlinear dynamics estimated from interaction traces. The spectrum yields three machine-checkable certificates: the sub-dominant eigenvalue $λ_2$ fixes the intrinsic timescale of reasoning and yields a convergence deadline computable \emph{before} the debate runs; its eigenvector names the coherent factions the collective reasons in, and $|λ_2|$ certifies when that explanation is valid; and the leading spectral coordinates form a compressed, auditable message basis. On an attention-consensus model, the deadline tracks observed convergence with log--log correlation $0.93$ and bounds it in 96\% of 24 configurations; attribution is exact whenever the spectrum certifies metastability; eight of 32 coordinates preserve the decision at 99.7\% fidelity; and a certificate learned from 15 debates held on 60/60 held-out debates. The study runs in minutes on a CPU, making spectral certification a practical layer for trustworthy collective reasoning.