Node-level Regression Uncertainty Quantification on Anaheim
100PICPER-GNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ER-GNN2026.05 | 100 | 2.09 | |
| BayesianNN2026.05 | 100 | 2.94 | |
| QpiGNNWidth penalty (lambda)=0.12026.05 | 99 | 0.74 | |
| QpiGNNWidth penalty (lambda)=opt.2026.05 | 93 | 0.4 | |
| QpiGNNWidth penalty (lambda)=0.52026.05 | 92 | 0.39 | |
| CF-GNN2026.05 | 90 | 3.22 | |
| CF-GNNoptimized=true2026.05 | 90 | 2.82 | |
| SQR-GNN2026.05 | 88 | 0.32 | |
| RQRadj.-GNN2026.05 | 85 | 0.5 | |
| MC Dropout2026.05 | 50 | 0.11 |