Semi-exact control functionals from Sard’s method

Summary A novel control variate technique is proposed for the post-processing of Markov chain Monte Carlo output, based on both Stein’s method and an approach to numerical integration due to Sard. The resulting estimators of posterior expected quantities of interest are proven to be polynomially exa...

Full description

Saved in:
Bibliographic Details
Published inBiometrika Vol. 109; no. 2; pp. 351 - 367
Main Authors South, L F, Karvonen, T, Nemeth, C, Girolami, M, Oates, C J
Format Journal Article
LanguageEnglish
Published Oxford University Press 01.06.2022
Subjects
Control variate
Stein operator
Variance reduction
Online AccessGet full text

Cover

More Information
Summary:Summary A novel control variate technique is proposed for the post-processing of Markov chain Monte Carlo output, based on both Stein’s method and an approach to numerical integration due to Sard. The resulting estimators of posterior expected quantities of interest are proven to be polynomially exact in the Gaussian context, while empirical results suggest that the estimators approximate a Gaussian cubature method near the Bernstein–von Mises limit. The main theoretical result establishes a bias-correction property in settings where the Markov chain does not leave the posterior invariant. Empirical results across a selection of Bayesian inference tasks are presented.
ISSN:0006-3444
1464-3510
DOI:10.1093/biomet/asab036