Gibbs Sampler

[jessegrabowski]

There's a scan-based implementation in the docs, but it's super cryptic. Scan is not necessary to implement this, we can just call the function in a loop. Still might be a nice place to start:

https://pytensor.readthedocs.io/en/latest/library/scan.html#using-shared-variables-gibbs-sampling

[ricardoV94]

Alternatively just a MH with the extra challenge of applying variable transforms automatically and replacing the separate value inputs by a raveled tensor

[lucianopaz]

I think a particle filter for tracking would be awesome. Something that shows continuous online learning, that pymc alone doesn’t really aim to tackle

[lucianopaz]

I'll leave this python implementation of a classic SIR particle filter for tracking that we could implement in pytensor