The scope of time series anomaly detection is increasingly shifting from univariate to multivariate contexts, as a growing number of real-world problems can no longer be adequately addressed by ...