Research map / Graph learning / Graph self-supervised learning
X-EviProbe: Post-hoc Parameter-Free Evidential Uncertainty Quantification for Frozen Graph Neural Networks
Chenghua Guo, Sihong Xie, Xi Zhang
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 latent geometry gate (ablation) | X-EviProbe · CiteSeer OOD detection | AUROC: 89.71 -> 98.19 | Improved |
| add latent geometry gate (ablation) | X-EviProbe · PubMed OOD detection | AUROC: 58.94 -> 83.30 | Improved |
| add evidence-strength propagation (ablation) | X-EviProbe · CiteSeer OOD detection | AUROC: 81.55 -> 98.19 | Improved |
| add evidence-strength propagation (ablation) | X-EviProbe · PubMed OOD detection | AUROC: 60.76 -> 83.30 | Improved |
| include validation nodes in reference set (ablation) | X-EviProbe · CiteSeer OOD detection | AUROC: 78.28 -> 98.19 | Improved |
| include validation nodes in reference set (ablation) | X-EviProbe · PubMed OOD detection | AUROC: 58.54 -> 83.30 | Improved |
| add logit magnitude evidence (ablation) | X-EviProbe · CiteSeer OOD detection | AUROC: 96.73 -> 98.19 | Improved |
| add logit magnitude evidence (ablation) | X-EviProbe · PubMed OOD detection | AUROC: 87.36 -> 83.30 | Hurt |
| use nearest neighbor instead of top-10 mean (ablation) | X-EviProbe · CiteSeer OOD detection | AUROC: 91.18 -> 98.19 | Improved |
| use nearest neighbor instead of top-10 mean (ablation) | X-EviProbe · PubMed OOD detection | AUROC: 70.37 -> 83.30 | Improved |
| use nearest neighbor instead of top-5 mean (ablation) | X-EviProbe · PubMed OOD detection | AUROC: 74.23 -> 83.30 | Improved |
| use Laplace kernel instead of Gaussian (ablation) | X-EviProbe · PubMed OOD detection | AUROC: 81.09 -> 83.30 | Improved |
Closest papers by meaning
- Energy-based Epistemic Uncertainty for Graph Neural Networks (NeurIPS 2024 · similarity 0.90)
- Accurate and Scalable Estimation of Epistemic Uncertainty for Graph Neural Networks (ICLR 2024 · similarity 0.88)
- Uncertainty Estimation for Heterophilic Graphs Through the Lens of Information Theory (ICML 2025 · similarity 0.88)
- Are Uncertainty Quantification Capabilities of Evidential Deep Learning a Mirage? (NeurIPS 2024 · similarity 0.87)
- Uncertainty Estimation by Flexible Evidential Deep Learning (NeurIPS 2025 · similarity 0.87)
- A Unified Uncertainty Representation for Graph Neural Networks via Doubly-Spectral Stochastic Expansion (NeurIPS 2026 · 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.
