Uncertainty Quantification on Gaussian Synthetic
0.04SharpnessMC dropout
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MC dropout2026.05 | 0.04 | 0.27 | |
| SQR-GNNBackbone=GraphSAGE2026.05 | 0.26 | 0.51 | |
| RQRadj.-GNNBackbone=GraphSAGE2026.05 | 0.29 | 0.54 | |
| QpiGNNlambda=0.5, Backbone=GraphSAGE2026.05 | 0.31 | 0.55 | |
| QpiGNNlambda=optimal, Backbone=GraphSAGE2026.05 | 0.42 | 0.65 | |
| QpiGNNlambda=0.1, Backbone=GraphSAGE2026.05 | 0.71 | 0.84 | |
| BayesianNN2026.05 | 8.92 | 2.98 | |
| CF-GNN2026.05 | 9.47 | 2.92 |