Counterfactual Graph Learning for Anomaly Detection with Feature Disentanglement and Generation (Student Abstract)

Graph anomaly detection has received remarkable research interests, and various techniques have been employed for enhancing detection performance. However, existing models tend to learn dataset-specific spurious correlations based on statistical associations. A well-trained model might suffer from performance degradation when applied to newly observed nodes with different environments. To handle this situation, we propose CounterFactual Graph Anomaly Detection model, CFGAD. In this model, we design a gradient-based separator to disentangle node features into class features and environment features. Then, we present a weight-varying diffusion model to combine class features and environment features from different nodes to generate counterfactual samples. These counterfactual samples will be adopted to enhance model robustness. Comprehensive experiments demonstrate the effectiveness of our CFGAD.

Authors

• Yutao Wei University of Electronic Science and Technology of China, Chengdu, Sichuan 610054, China
• Wenzheng Shu University of Electronic Science and Technology of China, Chengdu, Sichuan 610054, China
• Zhangtao Cheng University of Electronic Science and Technology of China
• Wenxin Tai University of Electronic Science and Technology of China, Chengdu, Sichuan 610054, China Kash Institute of Electronics and Information Industry, Kashgar 844000, China
• Chunjing Xiao Henan University, Kaifeng, Henan 475000, China
• Ting Zhong University of Electronic Science and Technology of China

Keywords

• Data Mining
• Graphical Models
• Knowledge Discovery