Latent feature reconstruction for unsupervised anomaly detection

Abstract Anomalies (or outliers) indicate a minority of data items that are quite different from the majority (inliers) dataset in certain aspect. Unsupervised anomaly detection (UAD) is an important but not yet extensively studied research topic. Recent deep learning based methods exploit reconstruction gap between inliers and outliers to discriminate them. However, it observed often decreases rapidly as training process goes. And there no reasonable way set stop point. To support effective UAD, we propose new UAD framework by introducing Latent Feature Reconstruction (LFR) layer can be applied recent methods. The LFR acts regularizer constrain latent features low-rank subspace which reconstructed well while cannot. We develop two implementing proposed with autoencoder architecture geometric transformation scheme. Experiments on five benchmarks show our achieve state-of-the-art performance most cases.