Include Aaron's Semantic Pointer element distribution in utils.dist.

[e2crawfo]

Useful for setting evaluation points.

[jgosmann]

Why did you close this issue?

Actually, I am planning to implement the analytic solution for this distribution (which is closely related to a Beta distribution).

[xchoo]

Reopening the issue. It would be extremely helpful if we had some way of determining the best distribution (be it using Jan's or Aaron's methods) for setting evaluation points (or intercepts) for SPA ensemble arrays.

[e2crawfo]

I closed it because I realized you can sample from it easily via UniformHypersphere...but yeah, it might be nice to have it included explicitly as a distribution, possibly including the PDF and CDF.

[jgosmann]

PDF/CDF will require scipy, but are definitely helpful if you do not want to distribute your evaluation points according to the distribution, but uniformly to cover the most likely x percent of inputs.

[jgosmann]

With #430 merged, is this still an issue?

[tbekolay]

I think this can be closed, yeah.