Dan Foreman-Mackey
Dan Foreman-Mackey
There aren't currently any plans to support the Type 3 transform for a few reasons: - I'm not totally sure of the use cases, - The logic will probably be...
@lgarrison and I have been chatting about the possibility of writing a paper describing what we're doing here. Something like "Differentiable programming with NUFFTs" or "NUFFTs for machine learning applications"...
Right now it's all just `assert`s. It would be good to have more informative error messages.
Currently, if any of the `finufft` methods fail with a non-zero error we just ignore it and keep on trucking. How should we handle this? It looks like the JAX...
I've derived the algorithms for matrix multiplication and solves, but I haven't been able to work out the factorization algorithm yet. There don't seem to be numerical issues for the...
This will require updating the backend to iterate over the batch dimension, but that shouldn't be too terribly hard. Then, we'd need to add a simple batching function. One question...
This should be possible using semi-separable matrices, but it'll take a little bit of reimplementation in the backend. Hopefully it's not too terrible!
I think that this must be possible using a forward and backward pass. At the very least, this must be possible when the predictive kernel is the same as the...
See https://github.com/exoplanet-dev/exoplanet-core/issues/2 for details.
Demo here: https://gist.github.com/dfm/01d4b172d5cc3b18f34143aeb2f8cb8e