A Compilation Target for Probabilistic Programming Languages
This work opens up a new research path for optimizing probabilistic programming systems, potentially benefiting developers and researchers in machine learning and AI, though it appears incremental as it builds on existing OS functionality.
The authors tackled the challenge of implementing forward inference techniques for probabilistic programming languages by creating a new intermediate representation called probabilistic C, which compiles to efficient machine code and links with OS libraries, resulting in an efficient, scalable, and portable compilation target.
Forward inference techniques such as sequential Monte Carlo and particle Markov chain Monte Carlo for probabilistic programming can be implemented in any programming language by creative use of standardized operating system functionality including processes, forking, mutexes, and shared memory. Exploiting this we have defined, developed, and tested a probabilistic programming language intermediate representation language we call probabilistic C, which itself can be compiled to machine code by standard compilers and linked to operating system libraries yielding an efficient, scalable, portable probabilistic programming compilation target. This opens up a new hardware and systems research path for optimizing probabilistic programming systems.