Stan Meeting, Oct 21 2018

Individual Reports:

Matthijs: Progress on OCaml parser for Stan3.

Ben G: RStanArm is ready to go, waiting on RStan next release. Worked on how to do includes of Stan programs in Stan programs.

Charles: Found class of models where the quadratic analyzer does poorly. Interfering with a model that Shosh wanted to fit. Going to post to discourse. Thinks this because of sharp discontinuities. Working on McGill slides.

Lauren: StanCon, hassles getting people to update their talks/slides.

Mitzi: Gotten through some pull requests. More on reference manual. Fully htmlized reference manual. Putting together a wiki page for how to add more documentation. Fixed bug in 2.18.

Breck: SGB admin stuff, proposal went out.

Bob: Wrote and distributed a Stan roadmap. Met with tensor flow probability (TFP). TFP ahead of us on matrix.

Jonah: Working on Andrew’s book. Getting ready for a release of Bayes Plot.

Paul: Keeping Ben and Mitzi busy. Worked on paper with Aki. Working on xval. Using ProdPred with BRMS.

Aki: Done paper with post-doc. Loo, gave Stan talk in Venice, in Marseille now.

Sean: GPU pull request for lower triangular ??. Avoiding copy, sharing dev ops knowledge. Trying to hire a free lancer to do mono-repo.


Discussion of deep copy bug. Turns out at most to be inefficient, not broken.

Discussion of how are we going to test the new parser.

Charles: Stochastic Gradients and HMC.

Bob: Stan Roadmap discussion.

1 Like

Are there any details online about this event? Who is Charles?

Charles is one of our PhD students, he is teaching a class at McGill for introduction to Stan in January I believe.


There are several accounts for “Charles”. This is a course, not an event?

@charlesm93, I think it is a class.


1 Like

I’m Charles. So far, the only Charles amongst the developers. I’ll be giving a two days workshop, this January. The event is still being planned, and a priori, I don’t think it’ll be open to the public. If it is, I’ll post something on the forum.

1 Like