Rethinking Feature Alignment in Generalist Graph Anomaly Detection: A Relational Fingerprint-based Approach
2026-05-25 • Machine Learning
Machine Learning
AI summaryⓘ
The authors developed a new method called ReFi-GAD to find unusual patterns in different kinds of graphs without needing to retrain the model for each new graph. Unlike previous methods that only tried to match features without understanding their meaning, ReFi-GAD uses a special 'Relational Fingerprint' that captures meaningful clues about anomalies by looking at both the context and structure of graphs. Their approach combines this fingerprint with a transformer model and a module that fine-tunes for each specific graph type. Tests on 14 different datasets showed that their method works better than existing ones.
Graph anomaly detectionGeneralist modelsRelational FingerprintTransformer encoderDomain adaptationFeature alignmentSemantic representationStructural contextPCANegative transfer
Authors
Yujing Liu, Yixin Liu, Yu Zheng, Alan Wee-Chung Liew, Xiaofeng Cao, Shirui Pan
Abstract
Generalist graph anomaly detection (GAD) aims to detect anomalies on unseen graphs without graph-specific retraining. Nevertheless, existing approaches primarily focus on aligning heterogeneous features across different data domains via PCA-based projection, which harmonizes feature dimensions ignores feature semantics. As a result, GAD models fail to learn transferable semantic knowledge, and even exhibit negative transfer on unseen graphs. To address this issue, we propose a Relational Fingerprint-based generalist GAD approach (ReFi-GAD for short), aligning heterogeneous raw features with a universal and semantics-aware Relational Fingerprint (ReFi) that encodes anomaly-indicative cues from both contextual and structural perspectives. Building on ReFi, we design a fingerprint-grounded generalist GAD model, which combines a transformer-based encoder to capture domain-invariant knowledge with an SNR-guided refinement module for domain-specific adaptation. Extensive experiments on 14 datasets demonstrate that ReFi-GAD significantly outperforms state-of-the-art methods.