Implement VonMises distribution for random angles

[fritzo]

No description provided.

[vishwakftw]

Will this not require an approximation / implementation of the Bessel function, which I think is not available in PyTorch? However, MKL and CUDA supports these as special functions, so I think a PR with Bessel would be required first.

[fritzo]

Gosh I haven't looked into the details of implementing .sample() and .log_prob(). Is the bessel function needed for sampling or for scoring?

[vishwakftw]

I believe it is required for scoring. Sampling can be done without Bessel functions.

[vishwakftw]

This is a good paper to read from https://www.jstor.org/stable/2346732?seq=1#page_scan_tab_contents.