Table 1 Performance boost of PINN on the Helmholtz benchmarks using transfer learning
From: Transfer learning for deep neural network-based partial differential equations solving
Subtask | Original | Transfer | Boost |
|---|---|---|---|
case 2 | 1.04e+00 | 5.71e-01 | 45.1% |
case 3 | 1.37e-01 | 7.34e-02 | 46.4% |
case 4 | 5.48e-01 | 1.54e-01 | 71.9% |
case 5 | 3.14e+00 | 5.55e-01 | 82.3% |
case 6 | 4.95e+00 | 2.21e+00 | 55.4% |
case 7 | 6.08e+00 | 5.02e+00 | 17.4% |