Deep Anomaly Detection with Outlier Exposure

TL;DR

This paper introduces Outlier Exposure, a method that trains anomaly detectors using auxiliary outlier datasets to improve detection of unseen anomalies in deep learning applications across vision and NLP tasks.

Contribution

It presents Outlier Exposure, a novel training approach leveraging auxiliary outlier data to enhance deep anomaly detection capabilities.

Findings

OE significantly improves anomaly detection performance.

OE mitigates likelihood-based issues in generative models.

Auxiliary dataset characteristics influence detection effectiveness.

Abstract

It is important to detect anomalous inputs when deploying machine learning systems. The use of larger and more complex inputs in deep learning magnifies the difficulty of distinguishing between anomalous and in-distribution examples. At the same time, diverse image and text data are available in enormous quantities. We propose leveraging these data to improve deep anomaly detection by training anomaly detectors against an auxiliary dataset of outliers, an approach we call Outlier Exposure (OE). This enables anomaly detectors to generalize and detect unseen anomalies. In extensive experiments on natural language processing and small- and large-scale vision tasks, we find that Outlier Exposure significantly improves detection performance. We also observe that cutting-edge generative models trained on CIFAR-10 may assign higher likelihoods to SVHN images than to CIFAR-10 images; we use OE to mitigate this issue. We also analyze the flexibility and robustness of Outlier Exposure, and identify characteristics of the auxiliary dataset that improve performance.