Mining In-distribution Attributes in Outliers for Out-of-distribution Detection

TL;DR

This paper introduces a novel OOD detection framework called MVOL that leverages the intrinsic in-distribution attributes present in outliers, improving detection accuracy by considering correlations often ignored in previous methods.

Contribution

The paper proposes a structured multi-view learning framework that explicitly models in-distribution attributes in outliers, providing theoretical insights and demonstrating superior performance over existing methods.

Findings

MVOL outperforms existing OOD detection methods in experiments.

It effectively utilizes auxiliary and noisy datasets for improved detection.

The framework offers theoretical validation of its effectiveness.

Abstract

Out-of-distribution (OOD) detection is indispensable for deploying reliable machine learning systems in real-world scenarios. Recent works, using auxiliary outliers in training, have shown good potential. However, they seldom concern the intrinsic correlations between in-distribution (ID) and OOD data. In this work, we discover an obvious correlation that OOD data usually possesses significant ID attributes. These attributes should be factored into the training process, rather than blindly suppressed as in previous approaches. Based on this insight, we propose a structured multi-view-based out-of-distribution detection learning (MVOL) framework, which facilitates rational handling of the intrinsic in-distribution attributes in outliers. We provide theoretical insights on the effectiveness of MVOL for OOD detection. Extensive experiments demonstrate the superiority of our framework to others. MVOL effectively utilizes both auxiliary OOD datasets and even wild datasets with noisy in-distribution data. Code is available at https://github.com/UESTC-nnLab/MVOL.