Elsa: Energy-based learning for semi-supervised anomaly detection

Abstract

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.

arXiv commentAccepted and published at BMVC2021

References in corpus (14)

• Distilling the Knowledge in a Neural Network
• A Simple Framework for Contrastive Learning of Visual Representations
• MCMC using Hamiltonian dynamics
• Popular Ensemble Methods: An Empirical Study
• Improved Regularization of Convolutional Neural Networks with Cutout
• FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
• Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks
• Energy-based Out-of-distribution Detection
• CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances
• Your Classifier is Secretly an Energy Based Model and You Should Treat it Like One
• Intriguing Properties of Contrastive Losses
• OCGAN: One-class Novelty Detection Using GANs with Constrained Latent Representations
• Adversarially Learned Anomaly Detection
• Novelty Detection Via Blurring