Enhancing Anomaly Detection Generalization through Knowledge Exposure: The Dual Effects of Augmentation

TL;DR

This paper introduces Knowledge Exposure, a novel anomaly detection method leveraging external knowledge and new testing protocols to improve out-of-distribution generalization, especially for samples with semantic-preserving transformations.

Contribution

It proposes a new testing protocol and the Knowledge Exposure method that uses external knowledge to better understand concept dynamics and improve anomaly detection under distribution shifts.

Findings

Superior performance on CIFAR-10, CIFAR-100, and SVHN datasets.

Effective differentiation of semantic shifts and transformations.

Enhanced generalization to out-of-distribution samples.

Abstract

Anomaly detection involves identifying instances within a dataset that deviate from the norm and occur infrequently. Current benchmarks tend to favor methods biased towards low diversity in normal data, which does not align with real-world scenarios. Despite advancements in these benchmarks, contemporary anomaly detection methods often struggle with out-of-distribution generalization, particularly in classifying samples with subtle transformations during testing. These methods typically assume that normal samples during test time have distributions very similar to those in the training set, while anomalies are distributed much further away. However, real-world test samples often exhibit various levels of distribution shift while maintaining semantic consistency. Therefore, effectively generalizing to samples that have undergone semantic-preserving transformations, while accurately detecting normal samples whose semantic meaning has changed after transformation as anomalies, is crucial for the trustworthiness and reliability of a model. For example, although it is clear that rotation shifts the meaning for a car in the context of anomaly detection but preserves the meaning for a bird, current methods are likely to detect both as abnormal. This complexity underscores the necessity for dynamic learning procedures rooted in the intrinsic concept of outliers. To address this issue, we propose new testing protocols and a novel method called Knowledge Exposure (KE), which integrates external knowledge to comprehend concept dynamics and differentiate transformations that induce semantic shifts. This approach enhances generalization by utilizing insights from a pre-trained CLIP model to evaluate the significance of anomalies for each concept. Evaluation on CIFAR-10, CIFAR-100, and SVHN with the new protocols demonstrates superior performance compared to previous methods.