Soss.jl: Probabilistic Metaprogramming with Julia

Probabilistic programming has been gaining popularity, as shown by recent releases of Google’s TensorFlow Probability, Facebook’s ProbTorch, and Uber’s Pyro. The Julia language has excellent support for numerical software development, in particular strong support for metaprogramming, which enables user-defined compiler-like transformations and optimizations. We’ll explore the capabilities and ideas behind Soss.jl, a new Julia library for probabilistic metaprogramming.
Who is this presentation for?
Practicing data scientists interested in quantifying uncertainty Anyone interested in Julia or metaprogramming
Prerequisite knowledge:
Basic understanding of probabilistic modeling (e.g., linear modeling, logisitic regression) Fundamental concepts of machine learning programming (e.g., numpy, sklearn)
What you'll learn?
Basic ideas in probabilistic programming Some benefits of Julia A glimpse into how probabilistic programming systems work

Dr. Chad Scherrer has been actively developing and using probabilistic programming systems since 2010, and served as technical lead for the language evaluation team in DARPA's Probabilistic Programming for Advancing Machine Learning ("PPAML") program. Much of his blog is devoted to describing Bayesian concepts using PyMC3, while his current Soss.jl project aims to improve execution performance by directly manipulating source code for models expressed in the Julia Programming Language . Dr. Scherrer is a Senior Data Scientist at Metis Seattle, where he teaches the Data Science Bootcamp.