Flex-GAD : Flexible Graph Anomaly Detection
Apu Chakraborty, Anshul Kumar, Gagan Raj Gupta
TL;DR
Flex-GAD addresses unsupervised node-level anomaly detection in attributed graphs by combining a community-aware GCN encoder with an attribute encoder, fused through a self-attention mechanism. It introduces a Smoothed Graph (Community-wise) to embed structural signals and uses a dual reconstruction objective with Jensen-Shannon divergence for stability. The approach achieves state-of-the-art average AUC across seven diverse datasets and substantially faster training compared with prior methods, due in part to adaptive encoder fusion and reduced hyperparameter tuning. Overall, Flex-GAD provides a robust, efficient solution that adapts to varying levels of feature homophily and graph structure, enabling practical deployment in complex networks.
Abstract
Detecting anomalous nodes in attributed networks, where each node is associated with both structural connections and descriptive attributes, is essential for identifying fraud, misinformation, and suspicious behavior in domains such as social networks, academic citation graphs, and e-commerce platforms. We propose Flex-GAD, a novel unsupervised framework for graph anomaly detection at the node level. Flex-GAD integrates two encoders to capture complementary aspects of graph data. The framework incorporates a novel community-based GCN encoder to model intra-community and inter-community information into node embeddings, thereby ensuring structural consistency, along with a standard attribute encoder. These diverse representations are fused using a self-attention-based representation fusion module, which enables adaptive weighting and effective integration of the encoded information. This fusion mechanism allows automatic emphasis of the most relevant node representation across different encoders. We evaluate Flex-GAD on seven real-world attributed graphs with varying sizes, node degrees, and attribute homogeneity. Flex-GAD achieves an average AUC improvement of 7.98% over the previously best-performing method, GAD-NR, demonstrating its effectiveness and flexibility across diverse graph structures. Moreover, it significantly reduces training time, running 102x faster per epoch than Anomaly DAE and 3x faster per epoch than GAD-NR on average across seven benchmark datasets.
