Graph Local Homophily Network for Anomaly Detection

In graph anomaly detection (GAD), the fact that anomalous nodes usually exhibit high heterophily, while most Graph Neural Networks (GNNs) have homophily assumptions, leads to poor performance. Many studies have attempted to solve this problem by employing a set of graph filters covering various frequencies. Their ultimate goal is to design the most appropriate spectral filter to capture the complex signals generated by normals and anomalies. The critical aspect lies in the fusion of information from filters with different frequency response functions. However, existing methods lack a clear indicator to guide the fusion of information at different frequencies.

In this paper, we find that local homophily is a valuable metric for assessing the weights of high- and low-frequency information at the node level, and we explicitly point out that the accuracy of local homophily is positively correlated with the accuracy of anomaly detection. Moreover, we unveil the phenomenon of camouflage in anomalous nodes, wherein these nodes disguise themselves by making their features resemble those of surrounding normals.

Based on this investigation, we propose the Graph Local Homophily Network for Anomaly Detection (GLHAD). Specifically, GLHAD first identifies the local homophily of nodes under the supervision of labeled nodes, where two contrasting paradigms are employed to resist the camouflage of anomalies. Then, a local homophily-based combination module combines low- and high-frequency signals based on the predicted local homophily. Finally, node representations from different layers are aggregated to make predictions. Comprehensive experiments on four anomaly detection datasets show that GLHAD outperforms other state-of-the-art baselines.

Ronghui Guo, Minghui Zou, Sai Zhang, Xiaowang Zhang, Zhizhi Yu, and Zhiyong Feng. (2024). "Graph Local Homophily Network for Anomaly Detection." Proceedings of the 33rd ACM International Conference on Information and Knowledge Management (CIKM '24), pp. 706–716.
Download Paper