Interpretable Graph-Level Anomaly Detection via Contrast with Normal Prototypes

TL;DR

ProtoGLAD is an unsupervised, interpretable graph anomaly detection framework that identifies anomalies by contrasting them with normal prototype graphs, offering both competitive detection and human-understandable explanations.

Contribution

It introduces a prototype-based approach that explicitly references real normal graphs for explanation, improving interpretability over existing methods.

Findings

Achieves competitive detection performance

Provides human-interpretable explanations

Discovers multiple normal prototypes iteratively

Abstract

The task of graph-level anomaly detection (GLAD) is to identify anomalous graphs that deviate significantly from the majority of graphs in a dataset. While deep GLAD methods have shown promising performance, their black-box nature limits their reliability and deployment in real-world applications. Although some recent methods have made attempts to provide explanations for anomaly detection results, they either provide explanations without referencing normal graphs, or rely on abstract latent vectors as prototypes rather than concrete graphs from the dataset. To address these limitations, we propose Prototype-based Graph-Level Anomaly Detection (ProtoGLAD), an interpretable unsupervised framework that provides explanation for each detected anomaly by explicitly contrasting with its nearest normal prototype graph. It employs a point-set kernel to iteratively discover multiple normal prototype graphs and their associated clusters from the dataset, then identifying graphs distant from all discovered normal clusters as anomalies. Extensive experiments on multiple real-world datasets demonstrate that ProtoGLAD achieves competitive anomaly detection performance compared to state-of-the-art GLAD methods while providing better human-interpretable prototype-based explanations.