Colin Carroll
Colin Carroll is a software engineer at Google Research. In this role he focuses Bayesian computation and research, and contributes to a number of open source libraries, including bayeux, TensorFlow Probability, PyMC, and ArviZ. He received his PhD in mathematics from Rice University, where he researched geometric measure theory.
Sessions
07-30
14:00
30min
The state of Bayesian workflows in JAX
Colin Carroll
This talk provides an overview to Bayesian workflows in the JAX ecosystem, and is aimed at both newcomers and experienced practitioners. We will look at the Bayesian workflow, from model specification and prior selection, to choices for inference like MCMC, optimization, and VI, to posterior analysis and validation. We will discuss practical ways to use popular libraries in building modular, efficient workflows.
Filmhouse