An Explainable GNN Framework for Component-Level Anomaly Diagnosis
2026-08-10 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors study how to find problems in complicated industrial systems made up of many sensors. Instead of just looking at which sensors show unusual readings, they focus on understanding how these sensors influence each other, because issues can come from disrupted relationships, not just broken sensors. They create a new method using Graph Neural Networks that helps find which parts of the system are actually causing problems, making it easier to explain system failures. Their tests showed this approach works well to find and rank the real faulty components.
Industrial processesMultivariate time seriesAnomaly detectionGraph Neural NetworksSensor networksSystem diagnosisInter-sensor influencesExplainabilityFault detection
Authors
Sena Ozgunay, Louise Travé-Massuyès, Jean-Michel Loubes, Raul Sena Ferreira
Abstract
Industrial processes are complex systems composed of multiple interacting sensors that generate multivariate time series (MTS). Detecting anomalies in such systems is critical for reliability and safety, yet understanding their origin is equally important. Existing Graph Neural Network (GNN)based methods for anomaly detection primarily focus on sensor-level deviations and either attribute anomalies directly to the deviating sensors. When diagnosis is attempted, generally, the most deviated sensor is identified as a root cause of a system fault. However, in many industrial systems, anomalies do not arise from faulty sensors but from disruptions in the influences governing the system dynamics. We propose an explainable GNN-based anomaly detection framework that shifts the perspective from sensor-level anomalies to component-level diagnosis, hypothesizing that anomalous measurements are symptoms of altered inter-sensor influences. Experiments show that the method effectively identifies and prioritizes the true faulty components, providing interpretable insights into system failures.