Research map / Graph learning / Graph self-supervised learning
Towards Self-Interpretable Graph-Level Anomaly Detection
Yixin Liu, Kaize Ding, Qinghua Lu, Fuyi Li, Leo Yu Zhang, Shirui Pan
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 |
|---|---|---|---|
| Use dual-hypergraph views rather than augmentation-based views (ablation) | SIGNET · BM-MS | AD-AUC: 63.26 -> 94.01 | Improved |
| Use dual-hypergraph views rather than augmentation-based views (ablation) | SIGNET · MNIST-0 | AD-AUC: 48.75 -> 83.25 | Improved |
| Use dual-hypergraph views rather than structural-property-based views (ablation) | SIGNET · BM-MS | AD-AUC: 92.79 -> 94.01 | Improved |
| Use dual-hypergraph views rather than structural-property-based views (ablation) | SIGNET · MNIST-0 | AD-AUC: 55.29 -> 83.25 | Improved |
| Use one shared extractor instead of two extractors with SKL divergence (ablation) | SIGNET · BM-MS | AD-AUC: 92.44 -> 94.01 | Improved |
| Use one shared extractor instead of two extractors with SKL divergence (ablation) | SIGNET · MNIST-0 | AD-AUC: 81.77 -> 83.25 | Improved |
| Use InfoNCE rather than the JS MI estimator (ablation) | SIGNET · BM-MS | AD-AUC: 75.21 -> 94.01 | Improved |
| Use InfoNCE rather than the JS MI estimator (ablation) | SIGNET · MNIST-0 | AD-AUC: 71.78 -> 83.25 | Improved |
| Use InfoNCE rather than the DV MI estimator (ablation) | SIGNET · BM-MS | AD-AUC: 74.87 -> 94.01 | Improved |
| Use InfoNCE rather than the DV MI estimator (ablation) | SIGNET · MNIST-0 | AD-AUC: 70.31 -> 83.25 | Improved |
| Apply SIGNET instead of OCGTL | BM-MN · BM-MN | AD-AUC: 99.49 -> 93.41 | Hurt |
| Apply SIGNET instead of OCGIN | MNIST-1 · MNIST-1 | AD-AUC: 98.25 -> 90.12 | Hurt |
Closest papers by meaning
- UniFORM: Towards Unified Framework for Anomaly Detection on Graphs (AAAI 2025 · similarity 0.89)
- UniGAD: Unifying Multi-level Graph Anomaly Detection (NeurIPS 2024 · similarity 0.89)
- Leveraging Diffusion Model as Pseudo-Anomalous Graph Generator for Graph-Level Anomaly Detection (ICML 2025 · similarity 0.88)
- Rethinking Reconstruction-based Graph-Level Anomaly Detection: Limitations and a Simple Remedy (NeurIPS 2024 · similarity 0.88)
- ARC: A Generalist Graph Anomaly Detector with In-Context Learning (NeurIPS 2024 · similarity 0.88)
- Self-Discriminative Modeling for Anomalous Graph Detection (ICML 2025 · 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.
