Theory and Modern Applications
Table 1 Initial computational parameters, (∗) part of parameter adjustment
From: Adaptive neural-domain refinement for solving time-dependent differential equations
Comp. parameter | Value |
|---|---|
hidden-layer neurons∗ | 5 |
initial weight values | 0 |
initial learning rate∗ | 1e-3 |
number of epochs | 1e5 |
training points (\(n_{{\mathrm{TP}}}\)) | 9 |
verification points (\(n_{{\mathrm{VP}}}\)) | 11 |
SCNF order (m) | 5 |
resize parameter (δ) | 0.5 |