AgentSoV:基于变异流的大语言模型多智能体系统虚拟感知
AgentSoV: Virtual Sensing for LLM Multi-Agent Systems via Stream of Variation
AI总结:
提出AgentSoV虚拟感知框架,将LLM多智能体系统建模为随机动态系统,支持故障归因等三种系统级分析,为MAS结构设计提供统一方案。
摘要:
大语言模型(LLM)多智能体系统(MAS)通过智能体交互解决复杂任务,但其可靠性难以分析,因为故障既取决于单个智能体,也取决于信息与错误在系统中的传播方式。现有方法大多将此视为事后故障归因,未为异构MAS架构提供通用模型,也不支持对信息影响与可观测性的更广泛分析。我们提出AgentSoV,这是一种虚拟感知框架,通过变异流(Stream of Variation,SoV)将MAS执行建模为随机动态系统。AgentSoV将异构轨迹映射为诊断阶段,从可观测输出构建虚拟测量值,并使用状态空间模型学习跨阶段变异传播。其表示在阶段依赖方面具有架构感知性,在分析语言方面具有架构无关性,支持链式、编排式和有向无环图(DAG)结构的MAS。无需重新执行MAS,AgentSoV支持三种系统级分析:故障归因识别异常变异的输入位置;阶段输入影响量化中间输入对下游结果的影响程度;虚拟传感器放置在监测预算下选择互补观测值。在多种MAS架构上的实验表明,该方法为故障定位、信息影响和可观测性提供了统一方案。AgentSoV将智能体执行重新建模为动态系统,而非仅作为文本检查,为MAS结构与组件设计提供了见解。
英文摘要:
Large language model (LLM) multi-agent systems (MASs) solve complex tasks through interacting agents, yet their reliability is difficult to analyze because failures depend on both individual agents and how information and errors propagate across the system. Existing approaches largely treat this as post-hoc failure attribution, but do not provide a common model across heterogeneous MAS architectures or support broader analysis of information influence and observability. We introduce AgentSoV, a virtual-sensing framework that models MAS execution as a stochastic dynamic system through Stream of Variation (SoV). AgentSoV maps heterogeneous trajectories into diagnostic stages, constructs virtual measurements from observable outputs, and learns cross-stage variation propagation with a state-space model. Its representation is architecture-aware in stage dependencies yet architecture-agnostic in analytical language, supporting chain-structured, orchestration-based, and directed acyclic graph (DAG)-structured MASs. Without re-executing the MAS, AgentSoV supports three system-level analyses: failure attribution identifies where abnormal variation enters; stage-input influence quantifies how strongly intermediate inputs affect downstream outcomes; and virtual-sensor placement selects complementary observations under a monitoring budget. Experiments across multiple MAS architectures demonstrate a unified approach to fault localization, information influence, and observability. AgentSoV reframes agentic execution as a dynamic system to model rather than merely text to inspect, providing insights for MAS structure and component design.
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。