Composing inference algorithms as program transformations
This work addresses the inefficiency in developing probabilistic inference methods for researchers and practitioners, though it is incremental as it builds on existing probabilistic programming systems.
The paper tackles the problem of manually coding probabilistic inference procedures for each model and algorithm by automatically generating them from models using modular program-to-program transformations. The result is inference procedures that achieve comparable accuracy and speed to other probabilistic programming systems on real-world problems.
Probabilistic inference procedures are usually coded painstakingly from scratch, for each target model and each inference algorithm. We reduce this effort by generating inference procedures from models automatically. We make this code generation modular by decomposing inference algorithms into reusable program-to-program transformations. These transformations perform exact inference as well as generate probabilistic programs that compute expectations, densities, and MCMC samples. The resulting inference procedures are about as accurate and fast as other probabilistic programming systems on real-world problems.