Diffusion based Semantic Outlier Generation via Nuisance Awareness for Out-of-Distribution Detection

TL;DR

This paper introduces SONA, a diffusion-based framework for generating challenging semantic outliers that improve out-of-distribution detection by focusing on semantic differences while maintaining nuisance similarity.

Contribution

The paper presents a novel diffusion model-guided approach, SONA, for generating semantic outliers with controllable nuisance features, enhancing OOD detection performance.

Findings

Achieves 88% AUROC on near-OOD datasets.

Surpasses baseline methods by approximately 6% in AUROC.

Effectively generates outliers with explicit semantic discrepancies.

Abstract

Out-of-distribution (OOD) detection, which determines whether a given sample is part of the in-distribution (ID), has recently shown promising results through training with synthetic OOD datasets. Nonetheless, existing methods often produce outliers that are considerably distant from the ID, showing limited efficacy for capturing subtle distinctions between ID and OOD. To address these issues, we propose a novel framework, Semantic Outlier generation via Nuisance Awareness (SONA), which notably produces challenging outliers by directly leveraging pixel-space ID samples through diffusion models. Our approach incorporates SONA guidance, providing separate control over semantic and nuisance regions of ID samples. Thereby, the generated outliers achieve two crucial properties: (i) they present explicit semantic-discrepant information, while (ii) maintaining various levels of nuisance resemblance with ID. Furthermore, the improved OOD detector training with SONA outliers facilitates learning with a focus on semantic distinctions. Extensive experiments demonstrate the effectiveness of our framework, achieving an impressive AUROC of 88% on near-OOD datasets, which surpasses the performance of baseline methods by a significant margin of approximately 6%.