Research map / Graph learning / Graph self-supervised learning
CLINIC: Towards High-quality Graph Out-Of-Distribution Detection
Yifan Wang, Haodong Zhang, Changhu Wang, Tao Ren, Dongjie Wang, Wei Ju, Chong Chen, Xian-Sheng Hua, Xiao Luo
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 decision boundary compression loss Lr to Lins+Lcon (ablation) | CLINIC · BZR+COX2 | AUC (%): 80.96 -> 96.32 | Improved |
| add decision boundary compression loss Lr to Lins+Lcon (ablation) | CLINIC · ClinTox+LIPO | AUC (%): 69.35 -> 71.31 | Improved |
| add contextual-aware concordance loss Lcon to Lins+Lr (ablation) | CLINIC · BZR+COX2 | AUC (%): 95.73 -> 96.32 | Improved |
| add contextual-aware concordance loss Lcon to Lins+Lr (ablation) | CLINIC · ClinTox+LIPO | AUC (%): 67.93 -> 71.31 | Improved |
| add perturbation-aware concordance loss Lins to Lcon+Lr (ablation) | CLINIC · BZR+COX2 | AUC (%): 95.78 -> 96.32 | Improved |
| add perturbation-aware concordance loss Lins to Lcon+Lr (ablation) | CLINIC · BBBP+BACE | AUC (%): 65.80 -> 84.21 | Improved |
| use Lr alone versus Lins alone (ablation) | CLINIC · BZR+COX2 | AUC (%): 80.57 -> 94.52 | Improved |
| combine two loss components instead of one (ablation) | CLINIC · ten ID/OOD pairs | AUC (%) | Improved |
| use contextual-affinity graph and twin concordance learning versus GOOD-D | CLINIC · ten ID/OOD pairs | average AUC: +1.16% over HGOE, +2.65% over GOOD-D | Improved |
| set k2 too small or too close to k1 in affinity graph (ablation) | CLINIC · Figure 3 dataset pairs | AUC | Hurt |
| use too many reciprocal neighbors k (ablation) | CLINIC · Figure 4 dataset pairs | AUC | Hurt |
| vary number of clusters C (ablation) | CLINIC · Figure 4 dataset pairs | AUC | No clear effect |
Closest papers by meaning
- Structural Entropy Guided Unsupervised Graph Out-Of-Distribution Detection (AAAI 2025 · similarity 0.89)
- Graph Out-of-Distribution Detection Goes Neighborhood Shaping (ICML 2024 · similarity 0.89)
- GOLD: Graph Out-of-Distribution Detection via Implicit Adversarial Latent Generation (ICLR 2025 · similarity 0.88)
- Revisiting Score Propagation in Graph Out-of-Distribution Detection (NeurIPS 2024 · similarity 0.87)
- GOODAT: Towards Test-Time Graph Out-of-Distribution Detection (AAAI 2024 · similarity 0.87)
- Harnessing Feature Resonance under Arbitrary Target Alignment for Out-of-Distribution Node Detection (NeurIPS 2025 · similarity 0.86)
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.
