WeiPer: OOD Detection using Weight Perturbations of Class Projections

TL;DR

WeiPer introduces weight perturbations in class projections to enhance out-of-distribution detection, achieving state-of-the-art results and providing theoretical and empirical insights into its effectiveness.

Contribution

The paper proposes a simple perturbation technique in the final layer that improves OOD detection and introduces a new distance-based method leveraging this representation.

Findings

Achieves state-of-the-art OOD detection performance

Effective especially when OOD samples are close to in-distribution data

Provides theoretical and empirical understanding of WeiPer's effectiveness

Abstract

Recent advances in out-of-distribution (OOD) detection on image data show that pre-trained neural network classifiers can separate in-distribution (ID) from OOD data well, leveraging the class-discriminative ability of the model itself. Methods have been proposed that either use logit information directly or that process the model's penultimate layer activations. With "WeiPer", we introduce perturbations of the class projections in the final fully connected layer which creates a richer representation of the input. We show that this simple trick can improve the OOD detection performance of a variety of methods and additionally propose a distance-based method that leverages the properties of the augmented WeiPer space. We achieve state-of-the-art OOD detection results across multiple benchmarks of the OpenOOD framework, especially pronounced in difficult settings in which OOD samples are positioned close to the training set distribution. We support our findings with theoretical motivations and empirical observations, and run extensive ablations to provide insights into why WeiPer works.