Steffen Ridderbusch

Noise parameter `Sigma` can't be recovered from `PosteriorGPs`

5

I noticed that once you have a `PosteriorGP`, it is no longer possible to recover the noise level that was given in the construction of the `FiniteGP`. My quick fix...

The current documentation via the Readme is fine, but fairly long, and somewhat incomplete. This setup should be moved to Documenter.jl, also rebuilding the performance plots when deploying on master,...

Some functions that are available in MKL are currently not defined, as they are slower than base, but "more accurate". This should be evaluated, first by adding the various commented...

MKL libraries are known to be hampered on AMD processors. There appears to be a undocumented environment flag to improve on that, for now, as seen here: https://www.pugetsystems.com/labs/hpc/How-To-Use-MKL-with-AMD-Ryzen-and-Threadripper-CPU-s-Effectively-for-Python-Numpy-And-Other-Applications-1637/ The VML...

Implement "Unifying View of Sparse Approximations ..." methods

8

As far as I can tell, this package currently implements a general approximation distribution `q`, which can be optimized "as a whole", which means as many optimization variables as there...