Recent work on probabilistic programming languages

Mar 6, 2017 12:00 AM

Who: Daniel Huang \
When: 16.00 {{ | date_to_long_string }}\
Where: 5128 (Grouproom) \
Title: {{ page.title }}

In this talk, we will present some of our recent work on probabilistic programming languages. In the first half of the talk, we will describe a semantics for these languages based on Type-2 computable distributions. Such an approach enables us to reason denotationally about probabilistic programs as well as in terms of sampling. In the second half, we will describe a compiler for a simple probabilistic programming language. The compiler uses a sequence of intermediate languages to gradually and successively transform a specification of a probabilistic model into a Markov Chain Monte Carlo inference algorithm for execution on the CPU or GPU.

Previous Talks

{: .t60 } {% include list-posts tag=‘csstalk’%}