Solving Nonlinear PDEs on Diffusion-reaction n = 64 (test)
0.0001Relative L2 ErrorBaldwinian-PINN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Baldwinian-PINNHardware=NVIDIA GeForce RTX 3090, Time phase=lifetime-learning (fine-tuning) phase2023.12 | 0.0001 | 0.58 | |
| RBF (multiquadric)Optimization=ADAM (500) + L-BFGS, Time phase=L-BFGS optimization only2023.12 | 0.0081 | 280 | |
| PolynomialOptimization=ADAM (500) + L-BFGS, Time phase=L-BFGS optimization only2023.12 | 0.0085 | 249 | |
| MultitaskOptimization=ADAM (500) + L-BFGS, Time phase=L-BFGS optimization only2023.12 | 0.009 | 243 | |
| LMCOptimization=ADAM (500) + L-BFGS, Time phase=L-BFGS optimization only2023.12 | 0.0091 | 302 | |
| CenterOptimization=ADAM (500) + L-BFGS, Time phase=L-BFGS optimization only2023.12 | 0.0095 | 426 | |
| RandomOptimization=ADAM (500) + L-BFGS, Time phase=L-BFGS optimization only2023.12 | 0.011 | 1,073 |