Research map / Label-efficient and robust learning / Semi-supervised learning
Normality Calibration in Semi-supervised Graph Anomaly Detection
Guolei Zeng, Hezhe Qiao, Guoguo Ai, Jinsong Guo, Guansong Pang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add ScoreDA and NormReg (full GraphNC) on top of the GGAD teacher | GGAD · Average over 6 datasets | AUROC: 0.7003 -> 0.7533 | Improved |
| add ScoreDA and NormReg (full GraphNC) on top of the GGAD teacher | GGAD · Average over 6 datasets | AUPRC: 0.2745 -> 0.3610 | Improved |
| add ScoreDA to the teacher (ablation) | GGAD · Tolokers | AUROC: 0.5340 -> 0.5811 | Improved |
| add NormReg to ScoreDA (student) after ScoreDA only (ablation) | OT+ScoreDA · Tolokers | AUROC: 0.5811 -> 0.6505 | Improved |
| add NormReg to ScoreDA on Amazon (ablation) | OT+ScoreDA · Amazon | AUROC: 0.9511 -> 0.9613 | Improved |
| add NormReg to ScoreDA on T-Finance (ablation) | OT+ScoreDA · T-Finance | AUPRC: 0.2566 -> 0.3667 | Improved |
| apply NormReg to labeled normal nodes only instead of all nodes (ablation) | GraphNC · T-Finance | AUPRC: 0.1971 -> 0.3667 | Improved |
| add NormReg directly to teacher pre-training without ScoreDA (ablation) | GGAD · Amazon | AUROC: 0.9443 -> 0.8956 | Hurt |
| add NormReg directly to teacher pre-training without ScoreDA (ablation) | GGAD · T-Finance | AUROC: 0.8228 -> 0.8455 | Improved |
| plug GraphNC into DOMINANT teacher | DOMINANT · Average over 6 datasets | AUROC: 0.6162 -> 0.7088 | Improved |
| plug GraphNC into OCGNN teacher | OCGNN · Average over 6 datasets | AUROC: 0.6265 -> 0.6956 | Improved |
| apply GraphNC to degraded GGAD teacher with anomalies injected into labeled normals | GGAD* · Average over 6 datasets | AUROC: 0.5008 -> 0.6411 | Improved |
Closest papers by meaning
- Open-Set Graph Anomaly Detection via Normal Structure Regularisation (ICLR 2025 · similarity 0.91)
- ARC: A Generalist Graph Anomaly Detector with In-Context Learning (NeurIPS 2024 · similarity 0.90)
- Generative Semi-supervised Graph Anomaly Detection (NeurIPS 2024 · similarity 0.90)
- Semi-supervised Graph Anomaly Detection via Robust Homophily Learning (NeurIPS 2025 · similarity 0.89)
- Graph Anomaly Detection as Dynamical Transport: Training-Free Scoring via Empirical Bayes (NeurIPS 2026 · similarity 0.88)
- UniGAD: Unifying Multi-level Graph Anomaly Detection (NeurIPS 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
