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

Achref Jaziri, Martin Rogmann, Martin Mundt, Visvanathan Ramesh
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:1216-1224, 2026.

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.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-jaziri26a, title = { Beyond Binary Out of Distribution Detection: Characterizing Distributional Shifts with Multi-Statistic Diffusion Trajectories }, author = {Jaziri, Achref and Rogmann, Martin and Mundt, Martin and Ramesh, Visvanathan}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {1216--1224}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/jaziri26a/jaziri26a.pdf}, url = {https://proceedings.mlr.press/v300/jaziri26a.html}, 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. } }
Endnote
%0 Conference Paper %T Beyond Binary Out of Distribution Detection: Characterizing Distributional Shifts with Multi-Statistic Diffusion Trajectories %A Achref Jaziri %A Martin Rogmann %A Martin Mundt %A Visvanathan Ramesh %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-jaziri26a %I PMLR %P 1216--1224 %U https://proceedings.mlr.press/v300/jaziri26a.html %V 300 %X 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.
APA
Jaziri, A., Rogmann, M., Mundt, M. & Ramesh, V.. (2026). Beyond Binary Out of Distribution Detection: Characterizing Distributional Shifts with Multi-Statistic Diffusion Trajectories . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:1216-1224 Available from https://proceedings.mlr.press/v300/jaziri26a.html.

Related Material