Detecting OOD Samples via Optimal Transport Scoring Function

TL;DR

This paper introduces OTOD, a novel post hoc OOD detection score based on optimal transport theory, which leverages features, logits, and softmax information to improve detection performance without extra training.

Contribution

The paper proposes OTOD, a new optimal transport-based scoring function for OOD detection that captures geometric cues in network representations and outperforms existing methods.

Findings

OTOD outperforms state-of-the-art GEN by 7.19% in mean FPR@95 on CIFAR-10.

OTOD outperforms GEN by 12.51% in mean FPR@95 on WideResNet-28.

Theoretical guarantees are provided for OTOD.

Abstract

To deploy machine learning models in the real world, researchers have proposed many OOD detection algorithms to help models identify unknown samples during the inference phase and prevent them from making untrustworthy predictions. Unlike methods that rely on extra data for outlier exposure training, post hoc methods detect Out-of-Distribution (OOD) samples by developing scoring functions, which are model agnostic and do not require additional training. However, previous post hoc methods may fail to capture the geometric cues embedded in network representations. Thus, in this study, we propose a novel score function based on the optimal transport theory, named OTOD, for OOD detection. We utilize information from features, logits, and the softmax probability space to calculate the OOD score for each test sample. Our experiments show that combining this information can boost the performance of OTOD with a certain margin. Experiments on the CIFAR-10 and CIFAR-100 benchmarks demonstrate the superior performance of our method. Notably, OTOD outperforms the state-of-the-art method GEN by 7.19% in the mean FPR@95 on the CIFAR-10 benchmark using ResNet-18 as the backbone, and by 12.51% in the mean FPR@95 using WideResNet-28 as the backbone. In addition, we provide theoretical guarantees for OTOD. The code is available in https://github.com/HengGao12/OTOD.