Running probabilistic programs backwards

DOI10.1007/978-3-662-46669-8_3MaRDI QIDQ2802430zbMATH OpenOpenAlexFDO

Authors Neil Toronto, Jay McCarthy, David van Horn

Publication date 26 April 2016

Published in Programming Languages and Systems (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1412.4053

zbMATH Keywords

probabilitysemanticsdomain-specific languages

Mathematics Subject Classification ID

Theory of programming languages (68N15) Probability in computer science (algorithm analysis, random structures, phase transitions, etc.) (68Q87) Semantics in the theory of computing (68Q55)

Abstract: Many probabilistic programming languages allow programs to be run under constraints in order to carry out Bayesian inference. Running programs under constraints could enable other uses such as rare event simulation and probabilistic verification---except that all such probabilistic languages are necessarily limited because they are defined or implemented in terms of an impoverished theory of probability. Measure-theoretic probability provides a more general foundation, but its generality makes finding computational content difficult. We develop a measure-theoretic semantics for a first-order probabilistic language with recursion, which interprets programs as functions that compute preimages. Preimage functions are generally uncomputable, so we derive an abstract semantics. We implement the abstract semantics and use the implementation to carry out Bayesian inference, stochastic ray tracing (a rare event simulation), and probabilistic verification of floating-point error bounds.

Recommendations

• Commutative semantics for probabilistic programming
• Semantics for probabilistic programming: higher-order functions, continuous distributions, and soft constraints
• Measure transformer semantics for Bayesian machine learning
• A lambda-calculus foundation for universal probabilistic programming
• Measure transformer semantics for Bayesian machine learning

Cited in

(7)

• Measure transformer semantics for Bayesian machine learning
• A provably correct sampler for probabilistic programs
• An application of computable distributions to the semantics of probabilistic programming languages
• Commutative semantics for probabilistic programming
• scientific article; zbMATH DE number 1692951 (Why is no real title available?)
• Semantics for probabilistic programming: higher-order functions, continuous distributions, and soft constraints
• A monad for probabilistic point processes

This page was built for publication: Running probabilistic programs backwards

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2802430)