<ul><li>The widespread adoption of digital services has increased the need for anomaly detection in IT operations.</li><li>A unifying framework for benchmarking unsupervised anomaly detection methods is introduced.</li><li>The problem of shifts in normal behaviors in AIOps scenarios is highlighted.</li><li>The proposed approach, Domain-Invariant VAE for Anomaly Detection (DIVAD), outperforms existing methods.</li></ul>

Unsupervised Anomaly Detection in Multivariate Time Series across Heterogeneous Domains

Discover more