Beyond Binary Out-of-Distribution Detection: Characterizing Distributional Shifts with Multi-Statistic Diffusion Trajectories

TL;DR

This paper introduces DISC, a diffusion-based method that characterizes distributional shifts with multi-statistic trajectories, enabling both improved OOD detection and classification of OOD types.

Contribution

The work presents DISC, a novel diffusion model-based approach that captures multi-dimensional statistical features for enhanced OOD detection and classification.

Findings

DISC matches or exceeds state-of-the-art OOD detection performance.

DISC can classify the type of OOD data, not just detect it.

Experiments on image and tabular data validate DISC's effectiveness.

Abstract

Detecting out-of-distribution (OOD) data is critical for machine learning, be it for safety reasons or to enable open-ended learning. However, beyond mere detection, choosing an appropriate course of action typically hinges on the type of OOD data encountered. Unfortunately, the latter is generally not distinguished in practice, as modern OOD detection methods collapse distributional shifts into single scalar outlier scores. This work argues that scalar-based methods are thus insufficient for OOD data to be properly contextualized and prospectively exploited, a limitation we overcome with the introduction of DISC: Diffusion-based Statistical Characterization. DISC leverages the iterative denoising process of diffusion models to extract a rich, multi-dimensional feature vector that captures statistical discrepancies across multiple noise levels. Extensive experiments on image and tabular benchmarks show that DISC matches or surpasses state-of-the-art detectors for OOD detection and, crucially, also classifies OOD type, a capability largely absent from prior work. As such, our work enables a shift from simple binary OOD detection to a more granular detection.