Detecting Perspective Shifts in Multi-agent Systems

2025-12-04

AI Summary

This study addresses the challenge of monitoring dynamic evolution of individual and collective behaviors in black-box multi-agent systems. We propose the Temporal Data Kernel Perspective Space (TDKPS) joint embedding framework, which employs low-dimensional representation learning and kernel methods to map agent response sequences across time steps into a unified, comparable feature space. Subsequently, we design novel hypothesis testing procedures tailored to detect behavioral changes at both agent-level and population-level granularities. To our knowledge, this constitutes the first principled, interpretable statistical inference paradigm for monitoring multi-agent behavioral dynamics. Experiments demonstrate robustness to hyperparameter choices; sensitivity and specificity are validated on simulated digital persona systems; and natural experiments confirm the framework’s ability to significantly detect behavioral anomalies strongly associated with real-world exogenous events.

Technology Category

• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Generative models augmented with external tools and update mechanisms (or extit{agents}) have demonstrated capabilities beyond intelligent prompting of base models. As agent use proliferates, dynamic multi-agent systems have naturally emerged. Recent work has investigated the theoretical and empirical properties of low-dimensional representations of agents based on query responses at a single time point. This paper introduces the Temporal Data Kernel Perspective Space (TDKPS), which jointly embeds agents across time, and proposes several novel hypothesis tests for detecting behavioral change at the agent- and group-level in black-box multi-agent systems. We characterize the empirical properties of our proposed tests, including their sensitivity to key hyperparameters, in simulations motivated by a multi-agent system of evolving digital personas. Finally, we demonstrate via natural experiment that our proposed tests detect changes that correlate sensitively, specifically, and significantly with a real exogenous event. As far as we are aware, TDKPS is the first principled framework for monitoring behavioral dynamics in black-box multi-agent systems -- a critical capability as generative agent deployment continues to scale.

Problem

• Detects behavioral changes in multi-agent systems
• Monitors agent dynamics across time using embeddings
• Provides tests for black-box system behavioral shifts

Innovation

• Jointly embeds agents across time using TDKPS
• Proposes novel hypothesis tests for behavioral change detection
• Monitors black-box multi-agent systems via principled framework