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

ARC: A Generalist Graph Anomaly Detector with In-Context Learning

About

Graph anomaly detection (GAD), which aims to identify abnormal nodes that differ from the majority within a graph, has garnered significant attention. However, current GAD methods necessitate training specific to each dataset, resulting in high training costs, substantial data requirements, and limited generalizability when being applied to new datasets and domains. To address these limitations, this paper proposes ARC, a generalist GAD approach that enables a ``one-for-all'' GAD model to detect anomalies across various graph datasets on-the-fly. Equipped with in-context learning, ARC can directly extract dataset-specific patterns from the target dataset using few-shot normal samples at the inference stage, without the need for retraining or fine-tuning on the target dataset. ARC comprises three components that are well-crafted for capturing universal graph anomaly patterns: 1) smoothness-based feature Alignment module that unifies the features of different datasets into a common and anomaly-sensitive space; 2) ego-neighbor Residual graph encoder that learns abnormality-related node embeddings; and 3) cross-attentive in-Context anomaly scoring module that predicts node abnormality by leveraging few-shot normal samples. Extensive experiments on multiple benchmark datasets from various domains demonstrate the superior anomaly detection performance, efficiency, and generalizability of ARC.

Yixin Liu, Shiyuan Li, Yu Zheng, Qingfeng Chen, Chengqi Zhang, Shirui Pan• 2024

Related benchmarks

TaskDatasetResultRank
Graph Anomaly DetectionAMAZON
AUROC79.08
109
Graph Anomaly DetectionREDDIT
AUROC60.94
106
Graph Anomaly DetectionBlogCatalog
AUROC0.7498
101
Graph Anomaly DetectionWeibo
AUROC89.76
99
Graph Anomaly DetectionFacebook
AUROC0.6769
75
Graph Anomaly DetectionPubmed
AUC85.62
65
Graph Anomaly Detectionquestions
AUPRC4.16
59
Graph Anomaly DetectionACM
AUPRC0.4079
54
Graph Anomaly DetectionWeibo (test)
AUPRC6.42e+3
51
Graph Anomaly DetectionReddit (test)
AUPRC4.2
51
Showing 10 of 62 rows

Other info

Code

Follow for update