Generative Adversarial Networks for Anomaly Detection: A Systematic Literature Review

Abstract

In numerous research areas, anomaly identification is a major problem. Identifying and properly classifying data as anomalous is a challenging task that is resolved in various manners over the years. Different approaches like traditional, supervised, unsupervised, and semi-supervised are used for anomaly detection. In the literature, various machine learning-based anomaly detection algorithms exist. It is challenging to choose one anomaly detection algorithm from the several available algorithms because each algorithm puts forward its good detection performance. In recent years, generative adversarial networks have shown remarkable results for anomaly classification. This paper aims to represent a systematic literature review of generative adversarial network-based approaches for anomaly detection and highlights their pros.