Stan Meeting Notes Jan 3, 2019


#1

Mitzi: PyCmdStan, work on Python interface. Working on Bayes workflow book. Keeping documentation going. Google has indexed user guide online. Search is working on git books, have PR for search mc-stan.org.

Jonah: Rstan pull requests, edits to long section of work flow document. Book progress.

Andrew: Nothing done in past two weeks.

Breck: StanCon 2019 continues firming up. No executed contracts yet.

Bob: Not much done. Working on intro to stats. Got soil carbon proposal written.

Lauren: Not much done. Working on some papers.

Aki: Not much done. Working on thinning. Working on Rhat paper, horse shoe.

Ben G: Working on BART and GPs with Jennifer on talk she is going. Gearing up for the Stan class at NYU.

Matthijs: Preparing talks for conference. Worked on offset multiplier in math. Worked on Stan compiler, worked more on discrete params project in Stan.

Sebastian: Working on parallel foreach in Stan math, can take any C++ functor that is parallel in execution will work. Alternative to MapRect, not a replacement.

General discussion:
Aki: Algo test bench development, ideas. Has students who could support infrastructure for testing against. Build a framework for others test against our models. Eval infrastructure with meta-data.

Mitzi: Look at performance tests CmdStan in repo, was stat comps benchmarks.

Tests will be run from R/Python.

ESS/sec metrics.

Version numbers on models.

Can replace exampleModels with this body of models.