Regression on UCI ENERGY (test)
-4.18Negative Log LikelihoodnGEM
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| nGEMOptimizer=Soap2026.02 | -4.18 | 0.3 | — | — | — | — | |
| NLLOptimizer=Soap2026.02 | -3.99 | 0.72 | — | — | — | — | |
| nGEMOptimizer=Adam2026.02 | -3.29 | 0.49 | — | — | — | — | |
| nGEMOptimizer=Muon2026.02 | -3.19 | 0.53 | — | — | — | — | |
| NLLOptimizer=Adam2026.02 | -3.09 | 0.95 | — | — | — | — | |
| NLLOptimizer=Muon2026.02 | -2.68 | 1.32 | — | — | — | — | |
| beta-NLLOptimizer=Adam2026.02 | -2.55 | 0.97 | — | — | — | — | |
| NLLOptimizer=KFAC2026.02 | -0.96 | 2.11 | — | — | — | — | |
| DWPdepth=32021.07 | -0.71 | — | — | — | — | — | |
| DWPdepth=22021.07 | -0.7 | — | — | — | — | — | |
| DWPdepth=42021.07 | -0.7 | — | — | — | — | — | |
| DWPdepth=52021.07 | -0.7 | — | — | — | — | — | |
| DGPdepth=22021.07 | -0.7 | — | — | — | — | — | |
| DGPdepth=32021.07 | -0.7 | — | — | — | — | — | |
| DGPdepth=42021.07 | -0.7 | — | — | — | — | — | |
| DGPdepth=52021.07 | -0.7 | — | — | — | — | — | |
| DIWPdepth=42021.07 | -0.7 | — | — | — | — | — | |
| nGEMOptimizer=KFAC2026.02 | -0.43 | 1.51 | — | — | — | — | |
| SDR-GP2026.05 | 0.37 | — | — | — | — | 0.05 | |
| DKL2026.05 | 0.45 | — | — | — | — | 0.09 | |
| UMAP-GP2026.05 | 0.53 | — | — | — | — | 0.13 | |
| Δ-UQtrials=20, train-test split=0.8-0.22022.07 | 0.56 | — | — | — | — | — | |
| DEnstrials=20, train-test split=0.8-0.22022.07 | 0.61 | — | — | — | — | — | |
| KSPCA-GP2026.05 | 0.67 | — | — | — | — | 0.22 | |
| NCA-GP2026.05 | 0.81 | — | — | — | — | 0.29 | |
| Bethe-FGinference_cost=single-pass2026.05 | 0.838 | — | — | — | — | — | |
| VBLLinference_cost=single-pass2026.05 | 0.852 | — | — | — | — | — | |
| DGP2026.05 | 0.88 | — | — | — | — | 0.12 | |
| SGHMCInference Type=Stochastic Gradient MCMC, Network Architecture=MLP (512-256-128), Activation=ReLU, Samples=1,0002023.02 | 0.92 | 0 | 2.71 | 4.1 | 195.98 | — | |
| BNNtrials=20, train-test split=0.8-0.22022.07 | 0.93 | — | — | — | — | — | |
| BPSInference Type=Piecewise Deterministic Markov Process, Network Architecture=MLP (512-256-128), Activation=ReLU, Samples=1,0002023.02 | 0.95 | 0 | 2.71 | 4.13 | 392.42 | — | |
| σBPSInference Type=Piecewise Deterministic Markov Process, Network Architecture=MLP (512-256-128), Activation=ReLU, Samples=1,0002023.02 | 0.95 | 0 | 2.73 | 4.15 | 800.66 | — | |
| Bethe-Obs2026.05 | 0.958 | — | — | — | — | — | |
| BoomerangInference Type=Piecewise Deterministic Markov Process, Network Architecture=MLP (512-256-128), Activation=ReLU, Samples=1,0002023.02 | 0.97 | 0 | 837.54 | 824.77 | 790.48 | — | |
| PBPtrials=20, train-test split=0.8-0.22022.07 | 1.01 | — | — | — | — | — | |
| VMG2021.02 | 1.06 | — | — | — | — | — | |
| VSD2021.02 | 1.06 | — | — | — | — | — | |
| Bethe-Diaginference_cost=single-pass2026.05 | 1.107 | — | — | — | — | — | |
| SLANG2021.02 | 1.12 | — | — | — | — | — | |
| MCDtrials=20, train-test split=0.8-0.22022.07 | 1.21 | — | — | — | — | — | |
| ETD-BAL-MomType=Score-free2026.06 | 1.217 | 0.772 | — | — | — | — | |
| VD2021.02 | 1.3 | — | — | — | — | — | |
| D.E2021.02 | 1.38 | — | — | — | — | — | |
| ETD-SR-EucType=Score-free2026.06 | 1.418 | 0.987 | — | — | — | — | |
| BBB2021.02 | 1.45 | — | — | — | — | — | |
| ETD-UB-MomType=Score-free2026.06 | 1.468 | 1.031 | — | — | — | — | |
| SGLDInference Type=Stochastic Gradient MCMC, Network Architecture=MLP (512-256-128), Activation=ReLU, Samples=1,0002023.02 | 1.6 | 0.01 | 2.96 | 5.74 | 1,000 | — | |
| Deep Ensembleinference_cost=multi-pass, ensemble_size=52026.05 | 1.613 | — | — | — | — | — | |
| Laplace-Fullinference_cost=single-pass2026.05 | 1.626 | — | — | — | — | — | |
| MAPinference_cost=single-pass2026.05 | 1.629 | — | — | — | — | — | |
| MCD2021.02 | 1.72 | — | — | — | — | — | |
| SVGDType=Baselines2026.06 | 1.756 | 0.761 | — | — | — | — | |
| MC Dropoutinference_cost=multi-pass, samples=502026.05 | 1.764 | — | — | — | — | — | |
| SGLDType=Baselines2026.06 | 1.991 | 1.226 | — | — | — | — | |
| ETD-SR-MomType=Score-guided2026.06 | 2.237 | 1.744 | — | — | — | — | |
| ETD-SR-Mom (IS)Type=Score-guided, Importance Sampling=true2026.06 | 2.254 | 1.681 | — | — | — | — | |
| ETD-SR-Euc (IS)Type=Score-guided, Importance Sampling=true2026.06 | 2.272 | 1.714 | — | — | — | — | |
| DLEMethod Class=ID-GVI2023.05 | 2.43 | — | — | — | — | — | |
| GP-RBFtype=non-parametric2026.05 | 2.532 | — | — | — | — | — | |
| DEMethod Class=ID-GVI2023.05 | 2.83 | — | — | — | — | — | |
| MCMCData Normalization=non-normalized2025.12 | 2.84 | — | — | — | — | — | |
| MNF2021.02 | 3.18 | — | — | — | — | — | |
| AGF-SVGDType=Baselines2026.06 | 3.462 | 5.389 | — | — | — | — | |
| HABNNData Normalization=non-normalized2025.12 | 3.81 | — | — | — | — | — | |
| DRLEMethod Class=ID-GVI2023.05 | 4.01 | — | — | — | — | — | |
| TAGIData Normalization=non-normalized2025.12 | 4.8 | — | — | — | — | — | |
| SVIData Normalization=non-normalized2025.12 | 5.38 | — | — | — | — | — | |
| KBNNData Normalization=non-normalized2025.12 | 8.24 | — | — | — | — | — | |
| PBPData Normalization=non-normalized2025.12 | 21.64 | — | — | — | — | — | |
| LaplaceData Normalization=non-normalized2025.12 | 4,324 | — | — | — | — | — | |
| DropoutData Normalization=non-normalized2025.12 | 15,558 | — | — | — | — | — | |
| MSEOptimizer=Adam2026.02 | — | 0.95 | — | — | — | — |