B1985
Title: Multivariable time series anomaly detection using heuristic spatial temporal graph neural network
Authors: Yu Jiang - The Chinese University of Hong Kong (China) [presenting]
Abstract: Anomaly detection for multivariable time series in cyberphysical systems is crucial for preventing system failures and ensuring safe production. The presence of strong coupling between system variables and propagation effects imparts pronounced spatial-temporal characteristics to anomalies. Designing an effective anomaly detection algorithm necessitates consideration of the coupling relationships, propagation directionality, and causal time delays among variables. We propose a heuristic spatial-temporal graph neural network for detecting anomalies in multivariate time series data. The performance of our model is verified using four public datasets. Our results highlight the advantages of utilizing a sparse directed graph structure for extracting system coupling characteristics.
