Understanding the limitations of self-supervised learning for tabular anomaly detection

Open in new window