Out-of-Distribution Detection with a Single Unconditional Diffusion Model

TL;DR

This paper presents DiffPath, a novel method using a single unconditional diffusion model to detect out-of-distribution samples across diverse tasks, challenging the need for multiple specialized models.

Contribution

The introduction of DiffPath, a technique that leverages a single diffusion model for effective OOD detection across various datasets, reducing the need for task-specific models.

Findings

DiffPath performs competitively with models trained specifically for each task.

The method effectively measures rate-of-change and curvature in diffusion paths for OOD detection.

Extensive experiments validate the versatility of a single diffusion model across multiple OOD tasks.

Abstract

Out-of-distribution (OOD) detection is a critical task in machine learning that seeks to identify abnormal samples. Traditionally, unsupervised methods utilize a deep generative model for OOD detection. However, such approaches require a new model to be trained for each inlier dataset. This paper explores whether a single model can perform OOD detection across diverse tasks. To that end, we introduce Diffusion Paths (DiffPath), which uses a single diffusion model originally trained to perform unconditional generation for OOD detection. We introduce a novel technique of measuring the rate-of-change and curvature of the diffusion paths connecting samples to the standard normal. Extensive experiments show that with a single model, DiffPath is competitive with prior work using individual models on a variety of OOD tasks involving different distributions. Our code is publicly available at https://github.com/clear-nus/diffpath.