Leveraging Diffusion Model as Pseudo-Anomalous Graph Generator for Graph-Level Anomaly Detection

Jinyu Cai, Yunhe Zhang, Fusheng Liu, See-Kiong Ng
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:6409-6425, 2025.

Abstract

A fundamental challenge in graph-level anomaly detection (GLAD) is the scarcity of anomalous graph data, as the training dataset typically contains only normal graphs or very few anomalies. This imbalance hinders the development of robust detection models. In this paper, we propose Anomalous Graph Diffusion (AGDiff), a framework that explores the potential of diffusion models in generating pseudo-anomalous graphs for GLAD. Unlike existing diffusion-based methods that focus on modeling data normality, AGDiff leverages the latent diffusion framework to incorporate subtle perturbations into graph representations, thereby generating pseudo-anomalous graphs that closely resemble normal ones. By jointly training a classifier to distinguish these generated graph anomalies from normal graphs, AGDiff learns more discriminative decision boundaries. The shift from solely modeling normality to explicitly generating and learning from pseudo graph anomalies enables AGDiff to effectively identify complex anomalous patterns that other approaches might overlook. Comprehensive experimental results demonstrate that the proposed AGDiff significantly outperforms several state-of-the-art GLAD baselines.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-cai25l, title = {Leveraging Diffusion Model as Pseudo-Anomalous Graph Generator for Graph-Level Anomaly Detection}, author = {Cai, Jinyu and Zhang, Yunhe and Liu, Fusheng and Ng, See-Kiong}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {6409--6425}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/cai25l/cai25l.pdf}, url = {https://proceedings.mlr.press/v267/cai25l.html}, abstract = {A fundamental challenge in graph-level anomaly detection (GLAD) is the scarcity of anomalous graph data, as the training dataset typically contains only normal graphs or very few anomalies. This imbalance hinders the development of robust detection models. In this paper, we propose Anomalous Graph Diffusion (AGDiff), a framework that explores the potential of diffusion models in generating pseudo-anomalous graphs for GLAD. Unlike existing diffusion-based methods that focus on modeling data normality, AGDiff leverages the latent diffusion framework to incorporate subtle perturbations into graph representations, thereby generating pseudo-anomalous graphs that closely resemble normal ones. By jointly training a classifier to distinguish these generated graph anomalies from normal graphs, AGDiff learns more discriminative decision boundaries. The shift from solely modeling normality to explicitly generating and learning from pseudo graph anomalies enables AGDiff to effectively identify complex anomalous patterns that other approaches might overlook. Comprehensive experimental results demonstrate that the proposed AGDiff significantly outperforms several state-of-the-art GLAD baselines.} }
Endnote
%0 Conference Paper %T Leveraging Diffusion Model as Pseudo-Anomalous Graph Generator for Graph-Level Anomaly Detection %A Jinyu Cai %A Yunhe Zhang %A Fusheng Liu %A See-Kiong Ng %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-cai25l %I PMLR %P 6409--6425 %U https://proceedings.mlr.press/v267/cai25l.html %V 267 %X A fundamental challenge in graph-level anomaly detection (GLAD) is the scarcity of anomalous graph data, as the training dataset typically contains only normal graphs or very few anomalies. This imbalance hinders the development of robust detection models. In this paper, we propose Anomalous Graph Diffusion (AGDiff), a framework that explores the potential of diffusion models in generating pseudo-anomalous graphs for GLAD. Unlike existing diffusion-based methods that focus on modeling data normality, AGDiff leverages the latent diffusion framework to incorporate subtle perturbations into graph representations, thereby generating pseudo-anomalous graphs that closely resemble normal ones. By jointly training a classifier to distinguish these generated graph anomalies from normal graphs, AGDiff learns more discriminative decision boundaries. The shift from solely modeling normality to explicitly generating and learning from pseudo graph anomalies enables AGDiff to effectively identify complex anomalous patterns that other approaches might overlook. Comprehensive experimental results demonstrate that the proposed AGDiff significantly outperforms several state-of-the-art GLAD baselines.
APA
Cai, J., Zhang, Y., Liu, F. & Ng, S.. (2025). Leveraging Diffusion Model as Pseudo-Anomalous Graph Generator for Graph-Level Anomaly Detection. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:6409-6425 Available from https://proceedings.mlr.press/v267/cai25l.html.

Related Material