Graph Neural Network-Based Anomaly Detection in Multivariate Time Series
This addresses the problem of detecting system faults or attacks in sensor data for domains like industrial monitoring, with incremental improvements in accuracy and explainability.
The paper tackles anomaly detection in multivariate time series by proposing a method that combines structure learning with graph neural networks to capture inter-sensor relationships and provide explainability, achieving more accurate detection than baselines on real-world sensor datasets.
Given high-dimensional time series data (e.g., sensor data), how can we detect anomalous events, such as system faults and attacks? More challengingly, how can we do this in a way that captures complex inter-sensor relationships, and detects and explains anomalies which deviate from these relationships? Recently, deep learning approaches have enabled improvements in anomaly detection in high-dimensional datasets; however, existing methods do not explicitly learn the structure of existing relationships between variables, or use them to predict the expected behavior of time series. Our approach combines a structure learning approach with graph neural networks, additionally using attention weights to provide explainability for the detected anomalies. Experiments on two real-world sensor datasets with ground truth anomalies show that our method detects anomalies more accurately than baseline approaches, accurately captures correlations between sensors, and allows users to deduce the root cause of a detected anomaly.