VISTA: Unsupervised 2D Temporal Dependency Representations for Time Series Anomaly Detection

Sinchee Chin, Fan Zhang, Xiaochen Yang +5

Time Series Anomaly Detection (TSAD) is essential for uncovering rare and potentially harmful events in unlabeled time series data. Existing methods are highly dependent on clean,…