Structural-Temporal Coupling Anomaly Detection with Dynamic Graph Transformer

TL;DR

This paper introduces a novel dynamic graph transformer model that captures structural-temporal coupling for anomaly detection in evolving graphs, outperforming existing methods across multiple datasets.

Contribution

The paper proposes a new architecture that integrates structural and temporal features at multiple levels using a dynamic graph transformer with 2D positional encoding, addressing a key gap in anomaly detection.

Findings

Outperforms state-of-the-art models on six datasets

Effectively captures structural-temporal interactions in dynamic graphs

Demonstrates robustness in real-world anomaly detection scenarios

Abstract

Detecting anomalous edges in dynamic graphs is an important task in many applications over evolving triple-based data, such as social networks, transaction management, and epidemiology. A major challenge with this task is the absence of structural-temporal coupling information, which decreases the ability of the representation to distinguish anomalies from normal instances. Existing methods focus on handling independent structural and temporal features with embedding models, which ignore the deep interaction between these two types of information. In this paper, we propose a structural-temporal coupling anomaly detection architecture with a dynamic graph transformer model. Specifically, we introduce structural and temporal features from two integration levels to provide anomaly-aware graph evolutionary patterns. Then, a dynamic graph transformer enhanced by two-dimensional positional encoding is implemented to capture both discrimination and contextual consistency signals. Extensive experiments on six datasets demonstrate that our method outperforms current state-of-the-art models. Finally, a case study illustrates the strength of our method when applied to a real-world task.