Active anomaly detection based on deep one-class classification

TL;DR

This paper introduces a novel active learning framework for deep one-class classification-based anomaly detection, improving sample selection and training methods to enhance detection performance with fewer labeled samples.

Contribution

It proposes a new query strategy based on an adaptive boundary and a semi-supervised training method using noise contrastive estimation for Deep SVDD.

Findings

The combined approach outperforms existing methods on seven datasets.

Adaptive boundary query strategy improves anomaly sample selection.

Semi-supervised training enhances model performance with limited labels.

Abstract

Active learning has been utilized as an efficient tool in building anomaly detection models by leveraging expert feedback. In an active learning framework, a model queries samples to be labeled by experts and re-trains the model with the labeled data samples. It unburdens in obtaining annotated datasets while improving anomaly detection performance. However, most of the existing studies focus on helping experts identify as many abnormal data samples as possible, which is a sub-optimal approach for one-class classification-based deep anomaly detection. In this paper, we tackle two essential problems of active learning for Deep SVDD: query strategy and semi-supervised learning method. First, rather than solely identifying anomalies, our query strategy selects uncertain samples according to an adaptive boundary. Second, we apply noise contrastive estimation in training a one-class classification model to incorporate both labeled normal and abnormal data effectively. We analyze that the proposed query strategy and semi-supervised loss individually improve an active learning process of anomaly detection and further improve when combined together on seven anomaly detection datasets.