Gibbs Flow for Approximate Transport with Applications to Bayesian Computation

DOI10.1111/RSSB.12404MaRDI QIDQ5087396OpenAlexFDO

Authors Jeremy Heng, Arnaud Doucet, Yvo Pokern

Publication date 11 July 2022

Published in Journal of the Royal Statistical Society Series B: Statistical Methodology (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1509.08787

zbMATH Keywords

Markov chain Monte Carlosequential Monte Carlopath samplingmass transportnormalising constants

Mathematics Subject Classification ID

Statistics (62-XX)

Recommendations

• Optimal transport and variational Bayesian inference
• Bayesian and variational Bayesian approaches for flows in heterogeneous random media
• Estimation and annealing for Gibbsian fields
• A multiscale strategy for Bayesian inference using transport maps
• Transport Monte Carlo: High-Accuracy Posterior Approximation via Random Transport
• Stochastic approximation algorithms for partition function estimation of Gibbs random fields
• A note on parametric Bayesian inference via gradient flows
• An efficient estimator for Gibbs random fields
• Conditional density estimation and simulation through optimal transport
• On the properties of variational approximations of Gibbs posteriors

Cited in

(10)

• An Invitation to Sequential Monte Carlo Samplers
• Inference via low-dimensional couplings
• Tensor rank reduction via coordinate flows
• Coupling techniques for nonlinear ensemble filtering
• Constructing sampling schemes via coupling: Markov semigroups and optimal transport
• Sequential Monte Carlo with transformations
• Transport Monte Carlo: High-Accuracy Posterior Approximation via Random Transport
• Multivariate feedback particle filter rederived from the splitting-up scheme
• The transport map computed by iterated function system
• On Large Lag Smoothing for Hidden Markov Models

This page was built for publication: Gibbs Flow for Approximate Transport with Applications to Bayesian Computation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5087396)