Solving Parameterized Time-Dependent Allen-Cahn PDE (10D)
0.0036Relative L1 ErrorDeepSplitting
Evaluation results
| Method | Links | |
|---|---|---|
| DeepSplittingParadigm=Retrain per μ2026.06 | 0.0036 | |
| DeepBSDEParadigm=Retrain per μ2026.06 | 0.0046 | |
| PRISM-SDGDParadigm=Zero-Shot ∀μ2026.06 | 0.0631 | |
| PRISM-STDEParadigm=Zero-Shot ∀μ2026.06 | 0.0637 |