Collaborative AI Agents and Critics for Fault Detection and Cause Analysis in Network Telemetry

2026-03-31

AI Summary

This work addresses the challenge of multimodal collaborative intelligence for fault detection, severity assessment, and root cause analysis in network telemetry by proposing a communication-efficient federated multi-agent framework that preserves the privacy of local cost functions. The framework employs multiple actors (task-performing agents) and critics (feedback-providing evaluators), coordinated through a central server without requiring direct communication among agents or critics. This design ensures that communication overhead scales linearly with the number of modalities and remains independent of the number of agents. By integrating multi-timescale stochastic approximation, classical machine learning, and generative AI foundation models, the system achieves privacy-preserving collaborative optimization. Empirical validation on network telemetry tasks demonstrates its effectiveness, with theoretical guarantees on the convergence of time-averaged active states.

Abstract

We develop algorithms for collaborative control of AI agents and critics in a multi-actor, multi-critic federated multi-agent system. Each AI agent and critic has access to classical machine learning or generative AI foundation models. The AI agents and critics collaborate with a central server to complete multimodal tasks such as fault detection, severity, and cause analysis in a network telemetry system, text-to-image generation, video generation, healthcare diagnostics from medical images and patient records, etcetera. The AI agents complete their tasks and send them to AI critics for evaluation. The critics then send feedback to agents to improve their responses. Collaboratively, they minimize the overall cost to the system with no inter-agent or inter-critic communication. AI agents and critics keep their cost functions or derivatives of cost functions private. Using multi-time scale stochastic approximation techniques, we provide convergence guarantees on the time-average active states of AI agents and critics. The communication overhead is a little on the system, of the order of $\mathcal{O}(m)$, for $m$ modalities and is independent of the number of AI agents and critics. Finally, we present an example of fault detection, severity, and cause analysis in network telemetry and thorough evaluation to check the algorithm's efficacy.

Problem

• fault detection
• cause analysis
• network telemetry
• severity assessment

Innovation

• collaborative AI agents
• multi-critic federated system
• multi-time scale stochastic approximation
• privacy-preserving cost optimization
• network telemetry fault analysis