Spatiotemporal Inference on Matérn-1-2 simulated 2D grid 1.0 (N=1024)
0.009MSE (y_gp vs y_hat)DeepRV-gMLP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DeepRV-gMLPESS(l)/sec=6.47 ± 2.44, Infer Time (s)=510 ± 177.81, Averaged across lengthscales (10, 30, 50)=true2025.03 | 0.009 | 0.26 | |
| INLAInfer Time (s)=7 ± 0.07, Averaged across lengthscales (10, 30, 50)=true2025.03 | 0.411 | 13.59 | |
| DeepRV-MLPESS(l)/sec=0.87 ± 0.36, Infer Time (s)=309 ± 52.97, Averaged across lengthscales (10, 30, 50)=true2025.03 | 4.895 | 12.91 | |
| Inducing PtsESS(l)/sec=8.10 ± 4.68, Infer Time (s)=566 ± 184.98, Averaged across lengthscales (10, 30, 50)=true2025.03 | 6.436 | 9.95 | |
| RFFESS(l)/sec=0.00 ± 0.00, Infer Time (s)=1,028 ± 175.29, Averaged across lengthscales (10, 30, 50)=true2025.03 | 9.735 | 15.72 | |
| PriorCVAEESS(l)/sec=11.98 ± 5.13, Infer Time (s)=231 ± 13.62, Averaged across lengthscales (10, 30, 50)=true2025.03 | 17.752 | 26.64 |