Deep Graph-level Anomaly Detection by Glocal Knowledge Distillation

TL;DR

This paper introduces a deep graph-level anomaly detection method that learns both local and global normal patterns through a novel knowledge distillation technique, significantly outperforming existing models on diverse datasets.

Contribution

The paper proposes a new GAD approach using joint random distillation of graph and node representations to capture comprehensive normal patterns.

Findings

Outperforms seven state-of-the-art models on 16 datasets

Effectively detects both local and global anomalies in graphs

Demonstrates robustness across diverse real-world domains

Abstract

Graph-level anomaly detection (GAD) describes the problem of detecting graphs that are abnormal in their structure and/or the features of their nodes, as compared to other graphs. One of the challenges in GAD is to devise graph representations that enable the detection of both locally- and globally-anomalous graphs, i.e., graphs that are abnormal in their fine-grained (node-level) or holistic (graph-level) properties, respectively. To tackle this challenge we introduce a novel deep anomaly detection approach for GAD that learns rich global and local normal pattern information by joint random distillation of graph and node representations. The random distillation is achieved by training one GNN to predict another GNN with randomly initialized network weights. Extensive experiments on 16 real-world graph datasets from diverse domains show that our model significantly outperforms seven state-of-the-art models. Code and datasets are available at https://git.io/GLocalKD.