Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Anomaly Detection in Dynamic Graphs via Transformer

About

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.

Yixin Liu, Shirui Pan, Yu Guang Wang, Fei Xiong, Liang Wang, Qingfeng Chen, Vincent CS Lee• 2021

Related benchmarks

TaskDatasetResultRank
Graph Anomaly DetectionAMAZON
AUROC65.02
109
Graph Anomaly DetectionREDDIT
AUROC50.6
106
Graph Anomaly DetectionBlogCatalog
AUROC0.6058
101
Graph Anomaly DetectionWeibo
AUROC41.08
99
Graph Anomaly DetectionFacebook
AUROC0.7088
75
Graph Anomaly DetectionPubmed
AUC70.09
65
Graph Anomaly Detectionquestions
AUPRC3.45
59
Graph Anomaly DetectionACM
AUPRC0.1816
54
Graph Anomaly DetectionCora--
50
Graph Anomaly DetectionYelpChi
AUROC52.41
49
Showing 10 of 27 rows

Other info

Follow for update