We identify two problems with Wingate et al. lightweight implementation technique for probabilistic programs. Simple examples demonstrate that common, what should be semantic-preserving program transformations drastically alter the program behavior. We briefly describe an alternative technique that does respect the program refactoring. There remains a question of how to be really, formally sure that the MCMC acceptance ratio is computed correctly.
Problems of the Lightweight Implementation of Probabilistic Programming
This entry was posted in Uncategorized and tagged poster. Bookmark the permalink.