Anomaly Detection With Multiple-Hypotheses Predictions

TL;DR

This paper introduces a multi-hypotheses autoencoder framework for anomaly detection that efficiently models normal data and identifies out-of-distribution samples, outperforming previous methods on benchmark datasets.

Contribution

It proposes a novel multi-hypotheses autoencoder with a discriminator to improve anomaly detection by modeling normal data distribution more effectively.

Findings

Up to 3.9% improvement on CIFAR-10

Reduced error from 6.8% to 1.5% on real anomaly detection task

Enforces diversity across hypotheses to prevent mode collapse

Abstract

In one-class-learning tasks, only the normal case (foreground) can be modeled with data, whereas the variation of all possible anomalies is too erratic to be described by samples. Thus, due to the lack of representative data, the wide-spread discriminative approaches cannot cover such learning tasks, and rather generative models, which attempt to learn the input density of the foreground, are used. However, generative models suffer from a large input dimensionality (as in images) and are typically inefficient learners. We propose to learn the data distribution of the foreground more efficiently with a multi-hypotheses autoencoder. Moreover, the model is criticized by a discriminator, which prevents artificial data modes not supported by data, and enforces diversity across hypotheses. Our multiple-hypothesesbased anomaly detection framework allows the reliable identification of out-of-distribution samples. For anomaly detection on CIFAR-10, it yields up to 3.9% points improvement over previously reported results. On a real anomaly detection task, the approach reduces the error of the baseline models from 6.8% to 1.5%.