JoeWingbermuehle

on YouTube ↗15m

" summary

A Scala DSL generates Stan code and runs it, claiming a type-safe Stan DSL that maps Scala data to Stan types, surfaces errors in the IDE, and exports models as plain Stan; practical tools include model and result caching and CSV/JSON import. Reparameterizing a Neal funnel shows many divergences before and far fewer after using a noncentered parameterization, and model composition and reusable transforms compile and run, though the project is early.

AI-written summary · about

" chapters (8)

1. 0:00Scala Stan intro
2. 1:27DSL design in Scala Stan
3. 3:20Running models and handling results
4. 4:29Fixing divergent transitions
5. 5:52Composing models in Scala Stan
6. 7:41Scala Stan features and limits
7. 9:29Future directions for Scala Stan
8. 10:38Q&A

" related talks

• 33mScala DSLs and Probabilistic ProgrammingJoe Wingbermuehle · Strange Loop 2018
• 28mStan: why does it exist? when is it useful? why do I use it?Daniel Lee · PyData Eindhoven 2020
• 21mScala DSL for ML Training Set StratificationShiva Chaitanya · Scale By the Bay 2017
• 29mMotivating Probabilistic ProgrammingRahul Chitturi · Scale By The Bay 2018
• 29mExploring Scala Tooling ecosystemJeferson David Ossa · Typelevel Summit Lausanne 2019
• 24mLet the Scala compiler work for youAdelbert Chang · Typelevel Summit NYC 2017
• 01h 37mHands-on Scala.MetaEugene Burmako · Scala World 2015
• 42mScala: The Unpredicted Lingua Franca for Data ScienceDean Wampler, Andy Petrella · Scala Days Berlin 2016