Fast and correct gradient-based optimisation for probabilistic programming via smoothing

DOI10.1007/978-3-031-30044-8_18MaRDI QIDQ6535252zbMATH OpenFDO

Authors Basim Khajwal, Chih-Hao Luke Ong, Dominik Wagner

Publication date 24 November 2023

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://doi.org/10.1007/978-3-031-30044-8_18

zbMATH Keywords

probabilistic programmingreparameterisation gradienttype systemsvalue semanticsvariational inference

Mathematics Subject Classification ID

Theory of programming languages (68N15) Other programming paradigms (object-oriented, sequential, concurrent, automatic, etc.) (68N19) Mathematical aspects of software engineering (specification, verification, metrics, requirements, etc.) (68N30) Stochastic programming (90C15) Applications of mathematical programming (90C90)

Recommendations

• Automatic differentiation variational inference
• Continualization of probabilistic programs with correction
• Laplacian smoothing gradient descent
• Densities of almost surely terminating probabilistic programs are differentiable almost everywhere
• scientific article; zbMATH DE number 6860839

Cites work

• A Convenient Category for Higher-Order Probability Theory
• A lambda-calculus foundation for universal probabilistic programming
• A provably correct sampler for probabilistic programs
• Borel structures for function spaces
• Commutative semantics for probabilistic programming
• Comparative smootheology
• Convex optimization algorithms
• Densities of almost surely terminating probabilistic programs are differentiable almost everywhere
• Discontinuous Optimization by Smoothing
• Gradient Convergence in Gradient methods with Errors
• scientific article; zbMATH DE number 193292 (Why is no real title available?)
• Machine learning. A probabilistic perspective
• Monte Carlo gradient estimation in machine learning
• Probabilistic coherence spaces are fully abstract for probabilistic PCF
• Probability theory. A comprehensive course
• Semantics for probabilistic programming: higher-order functions, continuous distributions, and soft constraints

Cited in

(1)

• GradInf: Gradient Estimation as Probabilistic Inference

This page was built for publication: Fast and correct gradient-based optimisation for probabilistic programming via smoothing

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6535252)