ResGCN: attention-based deep residual modeling for anomaly detection on attributed networks

• Published in: Machine learning Vol. 111; no. 2; pp. 519 - 541
• Main Authors: Pei, Yulong, Huang, Tianjin, van Ipenburg, Werner, Pechenizkiy, Mykola
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.02.2022; Springer Nature B.V
• Subjects: Anomalies; Artificial Intelligence; Artificial neural networks; Computer Science; Control; Fraud; Machine Learning; Mechatronics; Modelling; Natural Language Processing (NLP); Nodes; Nonlinearity; Robotics; Simulation and Modeling; Smoothing; Sparsity; Special Issue: Foundations of Data Science; Anomaly Detection; Graph Convolutional Network; Attributed Networks; Attention Mechanism
• ISSN: 0885-6125; 1573-0565
• DOI: 10.1007/s10994-021-06044-0

Effectively detecting anomalous nodes in attributed networks is crucial for the success of many real-world applications such as fraud and intrusion detection. Existing approaches have difficulties with three major issues: sparsity and nonlinearity capturing, residual modeling, and network smoothing. We propose Residual Graph Convolutional Network (ResGCN), an attention-based deep residual modeling approach that can tackle these issues: modeling the attributed networks with GCN allows to capture the sparsity and nonlinearity, utilizing a deep neural network allows direct residual ing from the input, and a residual-based attention mechanism reduces the adverse effect from anomalous nodes and prevents over-smoothing. Extensive experiments on several real-world attributed networks demonstrate the effectiveness of ResGCN in detecting anomalies.