Research map / Label-efficient and robust learning / OOD and anomaly detection
Improvements on Uncertainty Quantification for Node Classification via Distance Based Regularization
Russell Alan Hart, Linlin Yu, Yifei Lou, Feng Chen
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 |
|---|---|---|---|
| re-tune entropy regularization weight using validation cross-entropy (ablation) | GPN · Citeseer OOD detection | AUROC epistemic with propagation: 70.74 -> 73.75 | Improved |
| tune activation function (ReLU, Sigmoid, GELU) in addition (ablation) | GPN-CE · Citeseer OOD detection | AUROC epistemic with propagation: 73.75 -> 77.97 | Improved |
| add distance-based regularization (ablation) | GPN-CE-ACT · Citeseer OOD detection | AUROC epistemic with propagation: 77.97 -> 78.98 | Improved |
| re-tune entropy regularization weight using validation cross-entropy (ablation) | GPN · PubMed OOD detection | AUROC epistemic with propagation: 73.91 -> 78.32 | Improved |
| tune activation function in addition (ablation) | GPN-CE · PubMed OOD detection | AUROC epistemic with propagation: 78.32 -> 78.32 | No clear effect |
| add distance-based regularization (ablation) | GPN-CE-ACT · PubMed OOD detection | AUROC epistemic with propagation: 78.32 -> 81.76 | Improved |
| add distance-based regularization (ablation) | GPN-CE-ACT · PubMed OOD detection | AUPR epistemic with propagation: 74.11 -> 78.16 | Improved |
| re-tune entropy weight on validation CE (ablation) | GPN · PubMed | ID accuracy: 94.08 -> 93.84 | Hurt |
| use full method instead of GPN | GPN · CoraML misclassification detection | AUPR aleatoric: 49.51 -> 89.95 | Improved |
| use full method instead of GPN | GPN · CoraML misclassification detection | AUROC aleatoric: 81.19 -> 75.8 | Hurt |
| use full method instead of GPN | GPN · Citeseer misclassification detection | AUROC aleatoric: 75.89 -> 69.15 | Hurt |
| remove propagation layer (w/o network) | Ours · CoraML OOD detection | AUROC epistemic: 87.20 -> 76.12 | Hurt |
Closest papers by meaning
- A Unified Uncertainty Representation for Graph Neural Networks via Doubly-Spectral Stochastic Expansion (NeurIPS 2026 · similarity 0.88)
- Accurate and Scalable Estimation of Epistemic Uncertainty for Graph Neural Networks (ICLR 2024 · similarity 0.87)
- Uncertainty Quantification with the Empirical Neural Tangent Kernel (NeurIPS 2025 · similarity 0.86)
- Uncertainty Estimation for Heterophilic Graphs Through the Lens of Information Theory (ICML 2025 · similarity 0.86)
- Graph Neural Stochastic Diffusion for Estimating Uncertainty in Node Classification (ICML 2024 · similarity 0.86)
- SURE: SUrvey REcipes for building reliable and robust deep networks (CVPR 2024 · similarity 0.85)
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.
