What's new with Turing.jl and its ecosystem | Erlend Fjelde | JuliaCon 2024

What's new with Turing.jl and its ecosystem by Tor Erlend Fjelde
PreTalx: https://pretalx.com/juliacon2024/talk...
Slides: https://pretalx.com/media/juliacon202...

Turing.jl is a probabilistic programming language in Julia, i.e. a framework that let's you do Bayesian inference for your problem at hand without too much hassle.

Turing.jl is one of the older packages in Julia, with the first commits dating all the way back to when Julia was a mere v0.4. Since then it has seen a steady increase in use and adoption, with many interesting applications and contributions. And as Julia has matured and improved, so has Turing.jl.

In addition to development on Turing.jl itself, the TuringLang team has also been working on a number of other packages that feed into the Turing ecosystem with the aim of modularizing the ecosystem and making it easier to use and extend, both inside and outside of Turing.jl.

And so, as we just passed Julia v1.10 and internal whispers of the first major release (v1.0) of Turing.jl are forming, I thought it would be a good time to give an overview of the current state of Turing.jl and what the future holds. I'll also give an overview over the full ecosystem of packages that make up the Turing ecosystem and how this can be used to hook your code into the Turing ecosystem or simply use it to build your own stuff on top of. Finally, I'll showcase some success stories of people using Turing.jl to solve real-world problems.