Stan General Meeting September 6


#1

Notes from Stan Thursday meeting:

Andrew: Working on Stan book, extending invitation if others want to contribute to the book.

Breck: StanCon, we have governance in the form of the Stan Governing Body (SGB) and Technical Working Group (TWG).

Matthijs: Working on GLMs, hopefully will work well with gpus and scale in number of params.

Yuling: Pull request work on for Stan.

Sean: Working on benchmarking with Ben, GLMs. Working on gpu pull request. Matrix multiply work is merged. Working on matrix communication in gpu. Working on Stan3 parser.

Ben Goodrich: R packages loaded to cran, started teaching, giving tutorial at R medicine conference at Yale.

Paul: Working on papers.

Aki: Did StanCon, getting ready to teach, did some work on RHat.

Ben Bales: Working on Adjoint Jacobian stuff.

Bob: Working on Stan book/users guide. In Berlin, did meetup for Bayer, meet with Amazon. Possible funding sources via Amazon.

Mitzi: Pull request processing.

Lauren: StanCon work.

Sebastian: Played around with speeding up ODE solver.

Brian: Working with Turing language. Looking a StanArm.


#2

Dear Breck,

thank you for the summary of the meeting. When I stumbled about the information on Andrew and a Stan book, do you mean Andrew Gelman and a potential update of his book about Bayesian analysis? I have checked Gelman’s website and blog, and couldn’t find any related information. If I am not mistaken about Andrew, do you have any information about when the book may become available (+/-)?

Thank you for your reply.


#3

Book is here. It is also buried someplace in git that I don’t know about.

This is meant to be a community effort, I don’t quite know how that will be managed but if anyone can crowd source a great textbook, we can.

Breck


#4

I don’t know all the details, but here’s bit more info. The Stan book is practically the Stan manual minus language and function reference. Stan manual was not like other software manuals, but included huge amount of great text on Bayesian inference and computation. Stan manual is now divided in two parts, and one part is basis for Stan book which will be further developed to be useful as a standalone book. Stan manual is open collaborative project, and Stan book is, too. As there is some ongoing work to divide the manual and transform it from latex to bookdown (non-trivial), it has not yet been advertised yet. I assume that when the current works stabilizes, it will be advertised more, and I guess the management is same as for other parts of Stan software and documentation.