Community call notes 09/20

McCoy · September 24, 2020, 1:54pm

Just so we get broad coverage - if you took notes, you might share them here as well (@mohamed82008, @cscherrer, @phg)

Also, here are the collected links from the chat:

From Alex:

From Chad:

https://julialang.zulipchat.com/#narrow/stream/240884-soss.2Ejl/topic/Turing.20interface.20for.20sampling/near/199905425

From McCoy:

github.com/Keno/Compiler3.jl

Determine applicability of new interfaces for probabilistic programming

opened 04:05PM - 16 Sep 20 UTC

femtomc

@DhairyaLGandhi @Keno Hi all. I'm trying to determine the applicability of s…ome of this work to some of my interests in PP. I recently started exploring a technique which utilizes Mjolnir and a simple type system to perform an optimization at `@generated` expansion time. The optimization is dependent on the "semantics" of the inference operation which you are performing (which is currently expressed through a combination of type inference and a reaching analysis). [I've written a small abstract here.](https://femtomc.github.io/mrb_dynamic_specialization.pdf) My usage of Mjolnir is sort of funky - and I'm guessing the original use case is not quite like this. It almost seems like the new compiler work applies - but I really don't understand it yet, so I can't say with certainty. Nonetheless, I wanted to start this issue as a reference so that this topic might be explored (I know AD is being explored as we speak).

From Mohammed:

At one point, Alex also mentioned Hakaru - which is here: http://hakaru-dev.github.io/

mohamed82008 · September 24, 2020, 1:58pm

My notes can be found here https://github.com/TuringLang/Turing.jl/issues/1410.