Implement Marginalized RVs

[ricardoV94]

Working in pymc-devs/pymc#5438, I realized that the pattern that we need to match for Marginalized RVs is not too bad:

import aesara.tensor as at
from aeppl import joint_logprob

weights_rv = at.random.dirichlet([1, 1], name="weights")
components = at.stack([at.random.normal(name="c1"), at.random.normal(name="c2")], axis=-1)
mix_rv = at.sum(components * weights_rv, axis=-1)
mix_rv.name = "mix"

# weights can also just be a deterministic, only thing that matters is that it adds up to 1
weights_vv = weights_rv.clone() 
mix_vv = mix_rv.clone()

joint_logprob({weights_rv: weights_vv, mix_rv: mix_vv})

[brandonwillard]

Don't forget to link to the related issues: #21.

[brandonwillard]

Actually, that might just be a duplicate.

[ricardoV94]

Actually, that might just be a duplicate.

I think it's slightly different than that issue but I have to think about it a bit more.

[ricardoV94]

Nope, this is simply wrong... The original one is indeed what I had in mind