An Application of Computable Distributions to the Semantics of Probabilistic Programs
share
In this chapter, we explore how (Type-2) computable distributions can be used to give both distributional and (algorithmic) sampling semantics to probabilistic programs with continuous distributions. Towards this end, we first sketch an encoding of computable distributions in a fragment of Haskell. Next, we show how topological domains can be used to model the resulting PCF-like language. Throughout, we hope to emphasize the connection of such an approach with ordinary programming.
READ FULL TEXT