Anomaly Detection in Dynamic Graphs via Transformer

TL;DR

This paper introduces TADDY, a Transformer-based framework that effectively detects anomalies in dynamic graphs by encoding nodes' structural and temporal roles, outperforming existing methods on multiple datasets.

Contribution

The paper proposes a novel Transformer-based framework with a comprehensive node encoding strategy for anomaly detection in dynamic graphs, addressing key challenges of unattributed nodes and coupled spatial-temporal patterns.

Findings

TADDY outperforms state-of-the-art methods on six real-world datasets.

The framework effectively encodes structural and temporal node information.

Experimental results show significant improvement in anomaly detection accuracy.

Abstract

Detecting anomalies for dynamic graphs has drawn increasing attention due to their wide applications in social networks, e-commerce, and cybersecurity. Recent deep learning-based approaches have shown promising results over shallow methods. However, they fail to address two core challenges of anomaly detection in dynamic graphs: the lack of informative encoding for unattributed nodes and the difficulty of learning discriminate knowledge from coupled spatial-temporal dynamic graphs. To overcome these challenges, in this paper, we present a novel Transformer-based Anomaly Detection framework for DYnamic graphs (TADDY). Our framework constructs a comprehensive node encoding strategy to better represent each node's structural and temporal roles in an evolving graphs stream. Meanwhile, TADDY captures informative representation from dynamic graphs with coupled spatial-temporal patterns via a dynamic graph transformer model. The extensive experimental results demonstrate that our proposed TADDY framework outperforms the state-of-the-art methods by a large margin on six real-world datasets.