Class-wise Thresholding for Robust Out-of-Distribution Detection

TL;DR

This paper introduces a class-wise thresholding method to enhance the robustness of out-of-distribution detection in neural networks, effectively handling label shift and improving detection reliability.

Contribution

The authors propose a simple class-wise thresholding scheme applicable to existing OoD detection methods, addressing label shift sensitivity and improving robustness.

Findings

Improved OoD detection robustness under label shift

Class-wise thresholding maintains detection performance

Applicable to multiple existing OoD detection algorithms

Abstract

We consider the problem of detecting OoD(Out-of-Distribution) input data when using deep neural networks, and we propose a simple yet effective way to improve the robustness of several popular OoD detection methods against label shift. Our work is motivated by the observation that most existing OoD detection algorithms consider all training/test data as a whole, regardless of which class entry each input activates (inter-class differences). Through extensive experimentation, we have found that such practice leads to a detector whose performance is sensitive and vulnerable to label shift. To address this issue, we propose a class-wise thresholding scheme that can apply to most existing OoD detection algorithms and can maintain similar OoD detection performance even in the presence of label shift in the test distribution.