Translating Recursive Probabilistic Programs to Factor Graph Grammars

TL;DR

This paper introduces a semantics-preserving method to translate first-order probabilistic programs with conditionals and recursion into factor graph grammars, enabling more efficient inference without enumerating all possible graphs.

Contribution

It presents the first formal translation from recursive probabilistic programs with conditionals to factor graph grammars, enhancing inference efficiency.

Findings

Enables inference without enumerating all factor graphs

Preserves semantics during translation

Supports models with recursion and conditionals

Abstract

It is natural for probabilistic programs to use conditionals to express alternative substructures in models, and loops (recursion) to express repeated substructures in models. Thus, probabilistic programs with conditionals and recursion motivate ongoing interest in efficient and general inference. A factor graph grammar (FGG) generates a set of factor graphs that do not all need to be enumerated in order to perform inference. We provide a semantics-preserving translation from first-order probabilistic programs with conditionals and recursion to FGGs.