What should API look like for a marginal optimization algorithm?

“Gradient-based marginal optimization”. There’s an inactive Stan repo GitHub - stan-dev/gmo: Inference on marginal distributions using gradient-based optimization and iirc also a Stan branch somewhere. Bob has written about it in Tutorial on Monte Carlo EM and variants for MML and MMAP. Maybe @andrewgelman and @avehtari have more materials on it?

1 Like