Stan General Meeting, 11am EDT Thursday October 3, 2019

Meetings
1

Hangouts Link: https://meet.google.com/gzm-wmum-pfm

Instructions: Ask to attend in the hangouts interface and someone should let you in in the first 10 minutes of the meeting. Email breck@mc-stan.org if you have problems or want to attend the physical meeting in New York City.

Please add your agenda items in replies.

2

I would like to add the following to the agenda. Hopefully its not too “micro-managmenty” for the general meeting:

@tadej is adding broadcasting to normal_id_glm_lpdf() , bernoulli_logit_glm_lpmf() , poisson_log_glm_lpmf() and neg_binomial_2_log_glm_lpmf() and would also like to tackle unifying what kind of vector inputs should be allowed/supported. After he implements whatever will be the agreement I will add that to stanc3.

He has raised a question in this issue. Any feedback would be very welcome.

Unfortunately I am not 100% sure I will make it to the meeting. Feel free to drop the item if I will not make it. If you still discuss it and I will not be there, I would like to kindly ask if someone just drops a line here or on the issue if there was any consensus reached. Many thanks!

3

The integration of the Intel TBB on Windows with CmdStan requires the users to set the PATH variable. As an alternative to that we could generate additional .bat files to the .exe files. I would like to run this by the group how they like this idea to deal with the need to point Windows to the TBB dlls.

4

YouTube kickoff

5
  • Where do the units test for stanc3 live? Right now, it looks like they live in a different repo, i.e. stan, which is a problem for developing new code.
  • Can we share code to do SBC? Is it ok to run 100 chains and use 1 chain for each test? Code, or a framework they’ve used on the cluster (maybe specifically Columbia’s cluster).
6

License for CmdStanR

7

Commercial support devs reporting from NumFOCUS

8

parameters, transformed parameters, generated quantities and variables in Stan and interface outputs

9

Brief notes:

Steve: Some pull requests got done.

Sebastian: Working on integrating TBB building blocks.

Ben G: RstanArm resubmit, got RStan to compile with 2.20, review of McElreath book.

Ben Bales: No notes.

Bob: Funding stuff, parallel adaptation looks promising.

Rok: Worked on GLMs for new compiler.

Jonah: Papers, working on stuff for Andrew, CmdStanR, pondering license. Going for BSD.

Sean: Finishing up roadmap stuff. GPU work with Rok.

Lauren: Back from Austrialia.

Charles: Newton solver PR is merged into math. Working with cluster at Columbia.

Breck: Bunch of SGB work, admin.

Saban??: Working on case studies.

Aki: Submitted grant application for Bayesian workflow. Working on bunch of papers.

Paul: Working with Aki on software assisted Bayesain work flow.

====General Meeting=========

Youtube video for introduction to Stan YouTube channel by Maggie Lieu our YouTube manager. (https://www.youtube.com/watch?v=n_wJncctuSE&feature=youtu.be)

Rok :GLMs support row vectors or not?

Sebastian: TBB is ready for 2.21 but windows has issues with path problems.

Where to submit funding requests: Answer: board@mc-stan.org