A Multi-directional Meta-Learning Framework for Class-Generalizable Anomaly Detection

TL;DR

This paper introduces a multidirectional meta-learning framework for class-generalizable anomaly detection, effectively identifying unseen anomalies by leveraging limited labeled anomaly data and normal data.

Contribution

The paper proposes a novel multidirectional meta-learning algorithm that enhances generalization to unseen anomalies through a two-level optimization process.

Findings

Improved detection of unseen anomalies in experiments.

Effective use of limited anomaly labels for training.

Enhanced generalization over existing methods.

Abstract

In this paper, we address the problem of class-generalizable anomaly detection, where the objective is to develop a unified model by focusing our learning on the available normal data and a small amount of anomaly data in order to detect the completely unseen anomalies, also referred to as the out-of-distribution (OOD) classes. Adding to this challenge is the fact that the anomaly data is rare and costly to label. To achieve this, we propose a multidirectional meta-learning algorithm -- at the inner level, the model aims to learn the manifold of the normal data (representation); at the outer level, the model is meta-tuned with a few anomaly samples to maximize the softmax confidence margin between the normal and anomaly samples (decision surface calibration), treating normals as in-distribution (ID) and anomalies as out-of-distribution (OOD). By iteratively repeating this process over multiple episodes of predominantly normal and a small number of anomaly samples, we realize a multidirectional meta-learning framework. This two-level optimization, enhanced by multidirectional training, enables stronger generalization to unseen anomaly classes.