Compiling Markov Chain Monte Carlo Algorithms for Probabilistic Modeling
The job of a typical compiler is to convert a program written in some source language into assembly code. In contrast, a probabilistic programming language (PPL) compiler converts a specification of a probabilistic model into a statistical inference algorithm. Hence, the requirements of such a compiler are markedly different from those for a conventional one. In this paper, we describe a compiler for a restricted PPL that expresses Bayesian networks. In particular, we present a sequence of intermediate languages (ILs) that guide a compiler in gradually and successively refining a declarative specification of a probabilistic model into an executable Markov Chain Monte Carlo (MCMC) inference algorithm. The compilation strategy produces \emph{composable} MCMC algorithms for execution on a CPU or GPU.
Mon 19 JunDisplayed time zone: Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna change
10:50 - 12:30 | Learning and ProbabilisticPLDI Research Papers at Actes, Civil Engineering Chair(s): Swarat Chaudhuri Rice University | ||
10:50 25mTalk | DemoMatch: API Discovery from Demonstrations PLDI Research Papers Media Attached | ||
11:15 25mTalk | Similarity of Binaries through re-Optimization PLDI Research Papers | ||
11:40 25mTalk | Synthesizing Program Input Grammars PLDI Research Papers Osbert Bastani Stanford University, Rahul Sharma Microsoft Research, Alex Aiken Stanford University, Percy Liang Stanford University Media Attached | ||
12:05 25mTalk | Compiling Markov Chain Monte Carlo Algorithms for Probabilistic Modeling PLDI Research Papers Daniel Huang Harvard University, Jean-Baptiste Tristan Oracle Labs, Greg Morrisett Cornell University Media Attached | ||