Non-Parametric Outlier Synthesis

TL;DR

The paper introduces NPOS, a flexible non-parametric framework for generating artificial out-of-distribution data to improve detection of OOD samples without assuming specific distributional forms.

Contribution

It proposes a novel non-parametric outlier synthesis method that does not rely on distributional assumptions, enhancing OOD detection performance.

Findings

Outperforms existing OOD detection methods significantly.

Flexible synthesis approach applicable to various embedding distributions.

Mathematically interpretable as rejection sampling.

Abstract

Out-of-distribution (OOD) detection is indispensable for safely deploying machine learning models in the wild. One of the key challenges is that models lack supervision signals from unknown data, and as a result, can produce overconfident predictions on OOD data. Recent work on outlier synthesis modeled the feature space as parametric Gaussian distribution, a strong and restrictive assumption that might not hold in reality. In this paper, we propose a novel framework, Non-Parametric Outlier Synthesis (NPOS), which generates artificial OOD training data and facilitates learning a reliable decision boundary between ID and OOD data. Importantly, our proposed synthesis approach does not make any distributional assumption on the ID embeddings, thereby offering strong flexibility and generality. We show that our synthesis approach can be mathematically interpreted as a rejection sampling framework. Extensive experiments show that NPOS can achieve superior OOD detection performance, outperforming the competitive rivals by a significant margin. Code is publicly available at https://github.com/deeplearning-wisc/npos.