Spurious-Aware Prototype Refinement for Reliable Out-of-Distribution Detection

TL;DR

SPROD is a prototype-based method that refines class prototypes to improve out-of-distribution detection by reducing spurious correlation biases, achieving superior performance without extra data or tuning.

Contribution

The paper introduces SPROD, a post-hoc prototype refinement technique that explicitly mitigates spurious correlations in OOD detection, applicable across various models and datasets.

Findings

SPROD improves AUROC by 4.8% on average.

SPROD reduces FPR@95 by 9.4% on average.

Demonstrates superior performance across multiple challenging OOD datasets.

Abstract

Out-of-distribution (OOD) detection is crucial for ensuring the reliability and safety of machine learning models in real-world applications, where they frequently face data distributions unseen during training. Despite progress, existing methods are often vulnerable to spurious correlations that mislead models and compromise robustness. To address this, we propose SPROD, a novel prototype-based OOD detection approach that explicitly addresses the challenge posed by unknown spurious correlations. Our post-hoc method refines class prototypes to mitigate bias from spurious features without additional data or hyperparameter tuning, and is broadly applicable across diverse backbones and OOD detection settings. We conduct a comprehensive spurious correlation OOD detection benchmarking, comparing our method against existing approaches and demonstrating its superior performance across challenging OOD datasets, such as CelebA, Waterbirds, UrbanCars, Spurious Imagenet, and the newly introduced Animals MetaCoCo. On average, SPROD improves AUROC by 4.8% and FPR@95 by 9.4% over the second best.