Elsa: Energy-based learning for semi-supervised anomaly detection
This addresses a critical limitation in anomaly detection for scenarios with contaminated data, offering a robust solution that is incremental over existing contrastive learning methods.
The paper tackled the problem of anomaly detection when training data contains unknown anomalies, proposing Elsa, a semi-supervised approach that combines energy-based models with contrastive learning to achieve state-of-the-art performance in multiple contamination scenarios.
Anomaly detection aims at identifying deviant instances from the normal data distribution. Many advances have been made in the field, including the innovative use of unsupervised contrastive learning. However, existing methods generally assume clean training data and are limited when the data contain unknown anomalies. This paper presents Elsa, a novel semi-supervised anomaly detection approach that unifies the concept of energy-based models with unsupervised contrastive learning. Elsa instills robustness against any data contamination by a carefully designed fine-tuning step based on the new energy function that forces the normal data to be divided into classes of prototypes. Experiments on multiple contamination scenarios show the proposed model achieves SOTA performance. Extensive analyses also verify the contribution of each component in the proposed model. Beyond the experiments, we also offer a theoretical interpretation of why contrastive learning alone cannot detect anomalies under data contamination.