Exploiting Autoencoder's Weakness to Generate Pseudo Anomalies

TL;DR

This paper introduces a novel method that exploits autoencoders' tendency to reconstruct anomalies well by generating pseudo anomalies through adaptive noise, enhancing anomaly detection performance across multiple datasets.

Contribution

It proposes a new approach to generate pseudo anomalies by exploiting autoencoders' weaknesses, improving their ability to distinguish normal from anomalous data.

Findings

Improved anomaly detection accuracy on multiple datasets.

Effective generation of pseudo anomalies enhances discriminative capability.

Method is broadly applicable across different data types.

Abstract

Due to the rare occurrence of anomalous events, a typical approach to anomaly detection is to train an autoencoder (AE) with normal data only so that it learns the patterns or representations of the normal training data. At test time, the trained AE is expected to well reconstruct normal but to poorly reconstruct anomalous data. However, contrary to the expectation, anomalous data is often well reconstructed as well. In order to further separate the reconstruction quality between normal and anomalous data, we propose creating pseudo anomalies from learned adaptive noise by exploiting the aforementioned weakness of AE, i.e., reconstructing anomalies too well. The generated noise is added to the normal data to create pseudo anomalies. Extensive experiments on Ped2, Avenue, ShanghaiTech, CIFAR-10, and KDDCUP datasets demonstrate the effectiveness and generic applicability of our approach in improving the discriminative capability of AEs for anomaly detection.