FANFOLD: Graph Normalizing Flows-driven Asymmetric Network for Unsupervised Graph-Level Anomaly Detection

TL;DR

FANFOLD introduces an asymmetric graph neural network using normalizing flows and knowledge distillation to improve unsupervised graph-level anomaly detection, effectively distinguishing normal and abnormal graphs across diverse datasets.

Contribution

The paper proposes a novel asymmetric network with normalizing flows and knowledge distillation for enhanced unsupervised graph anomaly detection, addressing symmetry issues and feature instability.

Findings

FANFOLD outperforms 9 baseline methods on 15 datasets.

Normalizing flows effectively model the distribution of normal graphs.

Asymmetric design improves anomaly discrimination.

Abstract

Unsupervised graph-level anomaly detection (UGAD) has attracted increasing interest due to its widespread application. In recent studies, knowledge distillation-based methods have been widely used in unsupervised anomaly detection to improve model efficiency and generalization. However, the inherent symmetry between the source (teacher) and target (student) networks typically results in consistent outputs across both architectures, making it difficult to distinguish abnormal graphs from normal graphs. Also, existing methods mainly rely on graph features to distinguish anomalies, which may be unstable with complex and diverse data and fail to capture the essence that differentiates normal graphs from abnormal ones. In this work, we propose a Graph Normalizing Flows-driven Asymmetric Network For Unsupervised Graph-Level Anomaly Detection (FANFOLD in short). We introduce normalizing flows to unsupervised graph-level anomaly detection due to their successful application and superior quality in learning the underlying distribution of samples. Specifically, we adopt the knowledge distillation technique and apply normalizing flows on the source network, achieving the asymmetric network. In the training stage, FANFOLD transforms the original distribution of normal graphs to a standard normal distribution. During inference, FANFOLD computes the anomaly score using the source-target loss to discriminate between normal and anomalous graphs. We conduct extensive experiments on 15 datasets of different fields with 9 baseline methods to validate the superiority of FANFOLD.