Discrete (mixed) density estimators

Open michaeldeistler opened this issue 1 year ago • 4 comments

For discrete (mixed) parameters.

Can probably recycle a lot of code from MNLE.

Jan 15 '24 07:01 michaeldeistler

Yes, MNLE implements a mixed estimator with a categorical distribution. It would be nice to extend this to discrete flows as well.

Jan 15 '24 07:01 janfb

Related to #968

Feb 29 '24 17:02 gmoss13

happy to work on this during the hackathon

Mar 15 '24 10:03 coschroeder

The MNLE classes are now refactored to match the API of the other build functions, including z-scoring and embedding nets. Thus, in principle, one can now also use the MNLE setup for posterior inference. However, it allows only for a single discrete column in x (theta).

An autoregressive density estimator on top would be the solution, see #1112

Aug 23 '24 16:08 janfb

We now have discrete estimators for multiple dimensions, see #1269

and this issue will be fixed by #1362

Mar 13 '25 06:03 janfb

fixed by https://github.com/sbi-dev/sbi/pull/1362

Potentially one could open a new issue for mixed in the sense of some dimensions are estimated with one network and other with another network.

Mar 20 '25 16:03 dgedon