Graph-level Anomaly Detection via Hierarchical Memory Networks

TL;DR

This paper introduces Hierarchical Memory Networks (HimNet), a novel graph autoencoder-based approach that learns hierarchical node and graph memory modules to effectively detect both local and global anomalies in graph datasets.

Contribution

HimNet is the first method to jointly learn hierarchical memory modules for fine-grained and holistic anomaly detection in graphs, outperforming existing techniques.

Findings

HimNet significantly outperforms state-of-the-art methods on 16 real-world datasets.

HimNet is robust to anomaly contamination.

The approach effectively detects both locally and globally abnormal graphs.

Abstract

Graph-level anomaly detection aims to identify abnormal graphs that exhibit deviant structures and node attributes compared to the majority in a graph set. One primary challenge is to learn normal patterns manifested in both fine-grained and holistic views of graphs for identifying graphs that are abnormal in part or in whole. To tackle this challenge, we propose a novel approach called Hierarchical Memory Networks (HimNet), which learns hierarchical memory modules -- node and graph memory modules -- via a graph autoencoder network architecture. The node-level memory module is trained to model fine-grained, internal graph interactions among nodes for detecting locally abnormal graphs, while the graph-level memory module is dedicated to the learning of holistic normal patterns for detecting globally abnormal graphs. The two modules are jointly optimized to detect both locally- and globally-anomalous graphs. Extensive empirical results on 16 real-world graph datasets from various domains show that i) HimNet significantly outperforms the state-of-art methods and ii) it is robust to anomaly contamination. Codes are available at: https://github.com/Niuchx/HimNet.