Prakhar Sharma
Prakhar Sharma
hey I assume this is Euler-bernoulli beam. You can simply use the 4th order PDE instead of using the weak form.
Yes you can look at [variational PINNs](https://arxiv.org/abs/1912.00873). This is probably not implemented in DeepXDE. You need to use your own code.
If you are not using equations you should try techniques suitable for purely data-driven surrogates. DeepXDE employs DeepONets for parametric learning. You could also try Fourier neural operators which is...
I think domain decomposition is not supported by DeepXDE. You can try NVIDIA Modulus as they have a comprehensive library to handle such cases.
`on_boundary` will only work when you define the geometry using modules available in `dde.geometry` . Since you know the coordinates of the boundaries and from the plot i can see...
You can only use Neumann BC when you provide the normal vectors. In DeepXDE, the base geometry class has a function named `boundary_normal`, which stores the normal vectors. https://github.com/lululxvi/deepxde/blob/44f9324236dc45f865af73430b46d0dd5b2b9ebc/deepxde/geometry/geometry.py#L77 The...
Glad it worked. I have solved similar problems, where the boundary is complicated. Since your boundaries are straight lines, you could try using slope-intercept form of line: y=mx+c. You need...
For such complex geometry, you can't use filter function which is a combination of `on_boundary` and `np.isclose`. You could try writing a code from scratch. Here is a code I...
The question is fundamental to optimisers. While, second order optimisers are known for faster convergence, this is not generally true in cases where the loss lanscape contains [saddle points](https://stats.stackexchange.com/a/301728/362800).
No this is not implemented in DeepXDE and to be honest, LRA (learning rate annealing) isn't very effectively on many problems. Anyways it is implemented in NVIDIA Modulus. Personally, I...