A New Approach to Probabilistic Programming Inference
Abstract
We introduce and demonstrate a new approach to inference in expressive probabilistic programming languages based on particle Markov chain Monte Carlo. Our approach is simple to implement and easy to parallelize. It applies to Turing-complete probabilistic programming languages and supports accurate inference in models that make use of complex control flow, including stochastic recursion. It also includes primitives from Bayesian nonparametric statistics. Our experiments show that this approach can be more efficient than previously introduced single-site Metropolis-Hastings methods.
Cite
@article{arxiv.1507.00996,
title = {A New Approach to Probabilistic Programming Inference},
author = {Frank Wood and Jan Willem van de Meent and Vikash Mansinghka},
journal= {arXiv preprint arXiv:1507.00996},
year = {2015}
}
Comments
Updated version of the 2014 AISTATS paper (to reflect changes in new language syntax). 10 pages, 3 figures. Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, JMLR Workshop and Conference Proceedings, Vol 33, 2014