A Subspace Projection Approach to Autoencoder-based Anomaly Detection

TL;DR

This paper introduces HFR-AE, a novel autoencoder framework that projects inputs into a subspace to enhance high-fidelity reconstruction, significantly improving anomaly detection performance across various architectures.

Contribution

The paper proposes a subspace projection method for autoencoders to improve anomaly detection by increasing the reconstruction error gap between normal and anomalous data.

Findings

HFR-AE improves AUROC by up to 13.4% over vanilla AE.

The method is effective across different AE architectures.

Simulation results validate the effectiveness of the proposed approach.

Abstract

Autoencoder (AE) is a neural network (NN) architecture that is trained to reconstruct an input at its output. By measuring the reconstruction errors of new input samples, AE can detect anomalous samples deviated from the trained data distribution. The key to success is to achieve high-fidelity reconstruction (HFR) while restricting AE's capability of generalization beyond training data, which should be balanced commonly via iterative re-training. Alternatively, we propose a novel framework of AE-based anomaly detection, coined HFR-AE, by projecting new inputs into a subspace wherein the trained AE achieves HFR, thereby increasing the gap between normal and anomalous sample reconstruction errors. Simulation results corroborate that HFR-AE improves the area under receiver operating characteristic curve (AUROC) under different AE architectures and settings by up to 13.4% compared to Vanilla AE-based anomaly detection.