LFCS Seminars: 18 July 2018 - Mitchell Wand

We present a complete reasoning principle for contextual equivalence in an untyped probabilistic language. The language includes continuous (real-valued) random variables, conditionals, and scoring. It also includes recursion, since the standard  call-by-value fixpoint combinator is expressible.

 

We demonstrate the usability of our characterization by proving several equivalence schemas, including familiar facts from lambda calculus as well as results specific to probabilistic programming. In particular, we use it to prove that reordering the random draws in a probabilistic program preserves contextual equivalence. This allows us to show, for example, the contextual equivalence of:

 

let x = e1 in y = e2 in e0

and

let y = e2 in x = e1 in e0

 

(provided x does not occur free in e2 and y does not occur free in e1), even though e1 and e2 may have sampling and scoring effects)

 

(Joint work with Ryan Culpepper, Theophilos Giannakopoulos, and Andrew Cobb)