Proposing automatic differentiation for modeling
What efficiency can be introduced in stingray using automatic differentiation / variational inference for stochastic systems? Inherently the mathematical modeling using modern methods will be improved as compared to the current infrastructure of NumPy and Scipy libraries and will be easier to implement too but what are major drawbacks in the current design or to reframe it better: what impact do these modern statistical modelings bring to stingray?
Hi @rashmiraj137, what specific ideas did you have to enhance or replace a Stingray modeling tool? Is there a specific use case that is weirdly slow that we need to speed up? It would be great if you could also reference the items in this issue: https://github.com/StingraySoftware/stingray/issues/520
Closing this