Lena

Yes, what I meant by "histogram" was just a vector of counts. (In this case I think it was actually a continuous variable that was binned, but that seemed to...

https://github.com/bollwyvl/nosebook might be useful?

As the one who added pipits last summer, I don't think it's worth maintaining; doesn't really exemplify anything in particular.

> The operation we actually want for dirichlet multinomial accepts two > such trees of total size O(n), (the counts and the pseudo-counts) with > potentially repeated elements, and samples...

I think that at least suff_stat_categorical is worth adding; in addition to the suff-stat-ness, it could also take advantage of the fancy tree structure from #453 for asymptotically faster sampling....

One idea I had was an "adapter" that takes a made bernoulli procedure and makes a binomial out of it (similarly categorical to multinomial, optionally also bernoulli to geometric, I...

I don't think so. From reading the code, it looks like registerAEKernels just maintains a global set of nodes that have AEKernels, and any primitive infer hits all of them...

Option: have categorical propagate instead of absorbing changes to the list of input choices? Then it's the same (ish) as what happens with if-expressions: usually the output will be resampled...

Another place where this sort of thing might come up is hierarchical Dirichlet processes, although I think the practice of implementing them using CRP + memoization does the right thing...

Lite tries to reuse old indices, presumably so that unapply/reapply will give the same one, but even Lite's implementation has some unspecified behavior; it stores empty indices in a Python...