[Pl-seminar] 11/3 Seminar: Ryan Culpepper, Contextual Equivalence of Probabilistic Programs

[Daniel Patterson]

NUPRL Seminar presents

Ryan Culpepper
Northeastern University

12:00-1:30PM
Thursday, November 3rd, 2016
Room 366 WVH (http://www.ccs.neu.edu/home/wand/directions.html)

Contextual Equivalence of Probabilistic Programs

Abstract:

In this talk I introduce a probabilistic programming language with
continuous random variables and soft constraints, define contextual
equivalence for the language, and then present a logical relation that is
sound with respect to contextual equivalence.

The question of equivalence for probabilistic programs is interesting and
difficult because the meaning of a program is a measure defined by
integrating over the space of all of the program's possible evaluations.
Thus creating a workable logical relation requires proving integral
equalities involving intermediate measures that are ``equal enough'' but
still respect contextual equivalence.

Bio:

Ryan Culpepper is a research scientist at Northeastern University. He works
on probabilistic programming as well as extensible languages and tools for
building domain-specific languages. He received his PhD at Northeastern and
did a postdoc at the University of Utah.
-------------- next part --------------
HTML attachment scrubbed and removed