Towards Fair Graph Anomaly Detection: Problem, Benchmark Datasets, and Evaluation

TL;DR

This paper defines the Fair Graph Anomaly Detection problem, introduces real-world datasets from social media platforms with sensitive attributes and anomalies, and evaluates existing methods on fairness and accuracy trade-offs.

Contribution

It provides a formal problem definition, creates realistic benchmark datasets, and assesses current methods' performance and fairness in graph anomaly detection.

Findings

Datasets reveal differences from synthetic benchmarks.

Existing methods show trade-offs between accuracy and fairness.

Evaluation highlights challenges in fair anomaly detection.

Abstract

The Fair Graph Anomaly Detection (FairGAD) problem aims to accurately detect anomalous nodes in an input graph while avoiding biased predictions against individuals from sensitive subgroups. However, the current literature does not comprehensively discuss this problem, nor does it provide realistic datasets that encompass actual graph structures, anomaly labels, and sensitive attributes. To bridge this gap, we introduce a formal definition of the FairGAD problem and present two novel datasets constructed from the social media platforms Reddit and Twitter. These datasets comprise 1.2 million and 400,000 edges associated with 9,000 and 47,000 nodes, respectively, and leverage political leanings as sensitive attributes and misinformation spreaders as anomaly labels. We demonstrate that our FairGAD datasets significantly differ from the synthetic datasets used by the research community. Using our datasets, we investigate the performance-fairness trade-off in nine existing GAD and non-graph AD methods on five state-of-the-art fairness methods. Our code and datasets are available at https://github.com/nigelnnk/FairGAD