Variational Bayes with synthetic likelihood

DOI10.1007/S11222-017-9773-3zbMATH Open1384.65015arXiv1608.03069OpenAlexW2512401777WikidataQ62899349 ScholiaQ62899349MaRDI QIDQ1704030FDOQ1704030

Authors: Victor M. H. Ong, David J. Nott, Minh-Ngoc Tran, S. A. Sisson, C. C. Drovandi 

Publication date: 8 March 2018

Published in: Statistics and Computing (Search for Journal in Brave)

Abstract: Synthetic likelihood is an attractive approach to likelihood-free inference when an approximately Gaussian summary statistic for the data, informative for inference about the parameters, is available. The synthetic likelihood method derives an approximate likelihood function from a plug-in normal density estimate for the summary statistic, with plug-in mean and covariance matrix obtained by Monte Carlo simulation from the model. In this article, we develop alternatives to Markov chain Monte Carlo implementations of Bayesian synthetic likelihoods with reduced computational overheads. Our approach uses stochastic gradient variational inference methods for posterior approximation in the synthetic likelihood context, employing unbiased estimates of the log likelihood. We compare the new method with a related likelihood free variational inference technique in the literature, while at the same time improving the implementation of that approach in a number of ways. These new algorithms are feasible to implement in situations which are challenging for conventional approximate Bayesian computation (ABC) methods, in terms of the dimensionality of the parameter and summary statistic.

Full work available at URL: https://arxiv.org/abs/1608.03069

Recommendations

• Likelihood-free inference in high dimensions with synthetic likelihood
• Robust Bayesian synthetic likelihood via a semi-parametric approach
• Expectation propagation for likelihood-free inference
• Likelihood-free MCMC
• An extended empirical saddlepoint approximation for intractable likelihoods

zbMATH Keywords

variational Bayesapproximate Bayesian computationstochastic gradient ascentsynthetic likelihoods

Mathematics Subject Classification ID

Bayesian inference (62F15)

Cites Work

• An extended empirical saddlepoint approximation for intractable likelihoods
• Automatic differentiation variational inference
• Bayesian Synthetic Likelihood
• PRMLT
• Sampling and Bayes' Inference in Scientific Modelling and Robustness
• Title not available (Why is that?)
• Title not available (Why is that?)
• Pattern recognition and machine learning.
• The pseudo-marginal approach for efficient Monte Carlo computations
• Extending approximate Bayesian computation methods to high dimensions via a Gaussian copula model
• Likelihood-free inference via classification
• A comparison of inferential methods for highly nonlinear state space models in ecology and epidemiology
• A comparative review of dimension reduction methods in approximate Bayesian computation
• Bayesian optimization for likelihood-free inference of simulator-based statistical models
• Title not available (Why is that?)
• Explaining variational approximations
• Bayesian model comparison with un-normalised likelihoods
• Efficient implementation of Markov chain Monte Carlo when using an unbiased likelihood estimator
• Unbiased Estimation of Some Multivariate Probability Densities and Related Functions
• A Stochastic Approximation Method
• Pre-processing for approximate Bayesian computation in image analysis
• Approximate Bayesian computational methods
• Title not available (Why is that?)
• Likelihood-free Bayesian estimation of multivariate quantile distributions
• Fully simplified multivariate normal updates in non-conjugate variational message passing
• Title not available (Why is that?)
• Expectation propagation for likelihood-free inference
• On some properties of Markov chain Monte Carlo simulation methods based on the particle filter
• Title not available (Why is that?)
• Simple consistent estimators of stable distribution parameters
• Title not available (Why is that?)
• Large-scale machine learning with stochastic gradient descent
• Weighted quantile-based estimation for a class of transformation distributions.
• Fixed-form variational posterior approximation through stochastic linear regression
• Likelihood-free Bayesian inference for \(\alpha\)-stable models
• Gaussian variational approximation with sparse precision matrices

Cited In (18)

• Likelihood-free inference in high dimensions with synthetic likelihood
• Multifidelity multilevel Monte Carlo to accelerate approximate Bayesian parameter inference for partially observed stochastic processes
• Bayesian Deep Net GLM and GLMM
• Fast and accurate variational inference for models with many latent variables
• An approximate likelihood perspective on ABC methods
• Unbiased MLMC-based variational Bayes for likelihood-free inference
• Stochastic variational inference for GARCH models
• Bayesian Synthetic Likelihood
• Robust Bayesian synthetic likelihood via a semi-parametric approach
• Variational Bayes on manifolds
• Title not available (Why is that?)
• Likelihood-free approximate Gibbs sampling
• Approximating Bayes in the 21st century
• Variational Bayesian Filtering
• Bayesian Inference Using Synthetic Likelihood: Asymptotics and Adjustments
• Variational Bayes Estimation of Discrete-Margined Copula Models With Application to Time Series
• Efficient Bayesian Synthetic Likelihood With Whitening Transformations
• Likelihood-free inference with deep Gaussian processes

Uses Software

• GPS-ABC
• ADVI
• ADADELTA
• PRMLT

This page was built for publication: Variational Bayes with synthetic likelihood

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1704030)