Stabilizing Adversarially Learned One-Class Novelty Detection Using Pseudo Anomalies

TL;DR

This paper introduces a robust adversarial anomaly detection method that stabilizes training by using pseudo anomalies derived from generator states, improving detection performance across diverse datasets.

Contribution

The study proposes a novel discriminator role shift and a generator state-based pseudo anomaly creation to enhance training stability and detection accuracy.

Findings

Achieved state-of-the-art results on six diverse datasets.

Demonstrated stable training and improved anomaly detection performance.

Validated effectiveness across image, video, medical, and network security domains.

Abstract

Recently, anomaly scores have been formulated using reconstruction loss of the adversarially learned generators and/or classification loss of discriminators. Unavailability of anomaly examples in the training data makes optimization of such networks challenging. Attributed to the adversarial training, performance of such models fluctuates drastically with each training step, making it difficult to halt the training at an optimal point. In the current study, we propose a robust anomaly detection framework that overcomes such instability by transforming the fundamental role of the discriminator from identifying real vs. fake data to distinguishing good vs. bad quality reconstructions. For this purpose, we propose a method that utilizes the current state as well as an old state of the same generator to create good and bad quality reconstruction examples. The discriminator is trained on these examples to detect the subtle distortions that are often present in the reconstructions of anomalous data. In addition, we propose an efficient generic criterion to stop the training of our model, ensuring elevated performance. Extensive experiments performed on six datasets across multiple domains including image and video based anomaly detection, medical diagnosis, and network security, have demonstrated excellent performance of our approach.