Unsupervised Spatio-Temporal Anomaly Detection in Industrial Cyber-physical Systems

The author proposes MAD-Transformer for fine-grained anomaly detection in industrial systems, emphasizing temporal and spatial dependencies.