AgentSoV: Virtual Sensing for LLM Multi-Agent Systems via Stream of Variation

Key points

• 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.1
• AgentSoV maps heterogeneous trajectories into diagnostic stages, constructs virtual measurements from observable outputs, and learns cross-stage variation propagation with a state-space model.1
• Experiments across multiple MAS architectures demonstrate a unified approach to fault localization, information influence, and observability.1
• AgentSoV reframes agentic execution as a dynamic system to model rather than merely text to inspect, providing insights for MAS structure and component design.1

Sources (1)

• [1]AgentSoV: Virtual Sensing for LLM Multi-Agent Systems via Stream of Variation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 17, 07:48 PM

  We introduce AgentSoV, a virtual-sensing framework that models MAS execution as a stochastic dynamic system through Stream of Variation (SoV).

  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.

Extractive summary: sentences quoted from the sources.