Diffusion for Explainable Unsupervised Anomaly Detection

Abstract

Statistical anomaly detection is critical across various domains, including healthcare, finance, industry, and cybersecurity. While supervised methods often achieve high performance, the limited availability of labeled data requires effective unsupervised techniques. In this paper, we introduce Dataset Sampling Iterative Learning (DSIL), a novel iterative learning framework for unsupervised anomaly detection leveraging generative modeling with diffusion. Our approach progressively refines an unlabeled dataset by identifying and removing anomalies, effectively approximating a semi-supervised setup. We demonstrate the efficiency of our framework with Diffusion Time Estimation (DTE). Furthermore, it enables better explainability through a novel approach of noised-feature discovery. Extensive experiments against unsupervised methods on both synthetic and real-world datasets demonstrate improved state-of-the-art performance. Finally, we suggest a novel usage of existing metrics to evaluate the explainability of anomaly detection models.

Resources

Citation

@inproceedings{vincent2025diffusion,
  title={Diffusion for Explainable Unsupervised Anomaly Detection},
  author={Vincent, Elouan and Drean, Alexandre and Perez, Julien and Plantevit, Marc and Robardet, Celine},
  booktitle={IEEE International Conference on Data Science and Advanced Analytics (DSAA)},
  year={2025},
  organization={IEEE}
}