The correlated pseudomarginal method

DOI10.48550/ARXIV.1511.04992MaRDI QIDQ122164zbMATH OpenOpenAlexWikidataFDO

Authors George Deligiannidis, Arnaud Doucet, Michael K. Pitt, George Deligiannidis, Arnaud Doucet, Michael K. Pitt

Publication date 16 November 2015

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/1511.04992

zbMATH Keywords

Metropolis-Hastings algorithmparticle filterrandom-effects modelasymptotic posterior normalitycorrelated random numbersintractable likelihoodweak convergence

Mathematics Subject Classification ID

Bayesian inference (62F15) Monte Carlo methods (65C05) Central limit and other weak theorems (60F05)

Abstract: The pseudo-marginal algorithm is a popular variant of the Metropolis--Hastings scheme which allows us to sample asymptotically from a target probability density $pi$, when we are only able to estimate an unnormalized version of $pi$ pointwise unbiasedly. It has found numerous applications in Bayesian statistics as there are many scenarios where the likelihood function is intractable but can be estimated unbiasedly using Monte Carlo samples. Using many samples will typically result in averages computed under this chain with lower asymptotic variances than the corresponding averages that use fewer samples. For a fixed computing time, it has been shown in several recent contributions that an efficient implementation of the pseudo-marginal method requires the variance of the log-likelihood ratio estimator appearing in the acceptance probability of the algorithm to be of order 1, which in turn usually requires scaling the number $N$ of Monte Carlo samples linearly with the number $T$ of data points. We propose a modification of the pseudo-marginal algorithm, termed the correlated pseudo-marginal algorithm, which is based on a novel log-likelihood ratio estimator computed using the difference of two positively correlated log-likelihood estimators. We show that the parameters of this scheme can be selected such that the variance of this estimator is order $1$ as $N,Tightarrowinfty$ whenever $N/Tightarrow0$. By combining these results with the Bernstein-von Mises theorem, we provide an analysis of the performance of the correlated pseudo-marginal algorithm in the large $T$ regime. In our numerical examples, the efficiency of computations is increased relative to the standard pseudo-marginal algorithm by more than 20 fold for values of $T$ of a few hundreds to more than 100 fold for values of $T$ of around 10,000-20,000.

Recommendations

• The pseudo-marginal approach for efficient Monte Carlo computations
• Large-sample asymptotics of the pseudo-marginal method
• Subsampling MCMC -- an introduction for the survey statistician
• On the efficiency of pseudo-marginal random walk Metropolis algorithms
• Convergence properties of pseudo-marginal Markov chain Monte Carlo algorithms

Cited in

(40)

• Augmented pseudo-marginal Metropolis-Hastings for partially observed diffusion processes
• Subsampling MCMC -- an introduction for the survey statistician
• Efficient inference for stochastic differential equation mixed-effects models using correlated particle pseudo-marginal algorithms
• Bayesian parameter inference for partially observed stochastic differential equations driven by fractional Brownian motion
• Metropolis–Hastings via Classification
• Gaussian process enhanced semi-automatic approximate Bayesian computation: parameter inference in a stochastic differential equation system for chemotaxis
• Efficient Bayesian model choice for partially observed processes: with application to an experimental transmission study of an infectious disease
• A Statistical Recurrent Stochastic Volatility Model for Stock Markets
• Data-cloning SMC\(^2\): a global optimizer for maximum likelihood estimation of latent variable models
• Particle methods for stochastic differential equation mixed effects models
• A flexible particle Markov chain Monte Carlo method
• Hamiltonian Monte Carlo with energy conserving subsampling
• Accelerating Bayesian inference for stochastic epidemic models using incidence data
• Optimal scaling of random walk Metropolis algorithms using Bayesian large-sample asymptotics
• An adaptive MCMC method for Bayesian variable selection in logistic and accelerated failure time regression models
• Unbiased Markov chain Monte Carlo for intractable target distributions
• Subsampling sequential Monte Carlo for static Bayesian models
• Correlated pseudo-marginal schemes for time-discretised stochastic kinetic models
• Ensemble MCMC: accelerating pseudo-marginal MCMC for state space models using the ensemble Kalman filter
• Bayesian estimation of dynamic asset pricing models with informative observations
• cPseudoMaRg
• Bayesian Inference in Common Microeconometric Models With Massive Datasets by Double Marginalized Subsampling
• Computationally efficient Bayesian estimation of high-dimensional Archimedean copulas with discrete and mixed margins
• The Block-Poisson Estimator for Optimally Tuned Exact Subsampling MCMC
• scientific article; zbMATH DE number 4151591 (Why is no real title available?)
• Generalized Bayesian likelihood-free inference
• Accelerating inference for stochastic kinetic models
• Mixture copulas with discrete margins and their application to imbalanced data
• Statistic selection and MCMC for differentially private Bayesian estimation
• The pseudo-marginal approach for efficient Monte Carlo computations
• Efficient \(\mathrm{SMC}^2\) schemes for stochastic kinetic models
• Bayesian static parameter estimation for partially observed diffusions via multilevel Monte Carlo
• Advanced Multilevel Monte Carlo Methods
• A point mass proposal method for Bayesian state-space model fitting
• Approximating Bayes in the 21st century
• Computing Bayes: from then `til now
• Biased online parameter inference for state-space models
• Bayesian Inference Using Synthetic Likelihood: Asymptotics and Adjustments
• Smoothing with couplings of conditional particle filters
• Sequentially guided MCMC proposals for synthetic likelihoods and correlated synthetic likelihoods

This page was built for publication: The correlated pseudomarginal method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q122164)