TN-AutoRCA: Benchmark Construction and Agentic Framework for Self-Improving Alarm-Based Root Cause Analysis in Telecommunication Networks

2025-07-24

AI Summary

Root cause analysis (RCA) in telecom networks faces three key challenges: complex graph-structured dependencies among alarms, lack of large-scale real-world benchmarks, and insufficient interpretability of causal reasoning. To address these, this paper proposes the first autonomous, iterative multi-agent framework tailored for telecom alarm RCA. It integrates graph neural networks to model alarm propagation dynamics, reinforcement learning for sequential decision optimization, and collaborative multi-agent causal inference. We further construct and open-source the first large-scale, real-world RCA benchmark—comprising multi-vendor equipment, heterogeneous network topologies, and diverse fault patterns. Evaluated on production network data, our method improves root cause localization accuracy by 23.6% and reduces average response time by 41%, while demonstrating superior generalization and scalability across complex topologies. Our core innovations include pioneering the self-improving agent paradigm for telecom RCA and establishing a standardized evaluation framework for reproducible, rigorous assessment.

Technology Category

• Multiagent Systems: Agent Communication
• Reasoning under Uncertainty: Causality
• Machine Learning: Causal Learning

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Root Cause Analysis (RCA) in telecommunication networks is a critical task, yet it presents a formidable challenge for Artificial Intelligence (AI) due to its complex, graph-based reasoning requirements and the scarcity of realistic benchmarks.

Problem

• Lack of realistic benchmarks for telecom RCA
• Complex graph-based reasoning challenges AI
• Need self-improving alarm-based RCA solutions

Innovation

• Agentic framework for self-improving RCA
• Benchmark construction for telecommunication networks
• Graph-based reasoning for alarm analysis