Parallel Gaussian process surrogate Bayesian inference with noisy likelihood evaluations
From MaRDI portal
Abstract: We consider Bayesian inference when only a limited number of noisy log-likelihood evaluations can be obtained. This occurs for example when complex simulator-based statistical models are fitted to data, and synthetic likelihood (SL) method is used to form the noisy log-likelihood estimates using computationally costly forward simulations. We frame the inference task as a sequential Bayesian experimental design problem, where the log-likelihood function is modelled with a hierarchical Gaussian process (GP) surrogate model, which is used to efficiently select additional log-likelihood evaluation locations. Motivated by recent progress in the related problem of batch Bayesian optimisation, we develop various batch-sequential design strategies which allow to run some of the potentially costly simulations in parallel. We analyse the properties of the resulting method theoretically and empirically. Experiments with several toy problems and simulation models suggest that our method is robust, highly parallelisable, and sample-efficient.
Recommendations
- Adaptive Gaussian Process Approximation for Bayesian Inference with Expensive Likelihood Functions
- A parallel Bayesian optimization method based on adaptive surrogate model
- Bayesian optimization for likelihood-free inference of simulator-based statistical models
- Bayesian design of experiments for intractable likelihood models using coupled auxiliary models and multivariate emulation
- Surrogate-based sequential Bayesian experimental design using non-stationary Gaussian processes
Cites work
- A supermartingale approach to Gaussian process based sequential design of experiments
- A tutorial on approximate Bayesian computation
- Accelerating Bayesian Synthetic Likelihood With the Graphical Lasso
- Accelerating pseudo-marginal MCMC using Gaussian processes
- Adaptive Gaussian Process Approximation for Bayesian Inference with Expensive Likelihood Functions
- An analysis of approximations for maximizing submodular set functions—I
- Approximate Bayesian computational methods
- Bayes-Hermite quadrature
- Bayesian calibration of computer models. (With discussion)
- Bayesian optimization for likelihood-free inference of simulator-based statistical models
- Bayesian Probabilistic Numerical Methods
- Bayesian Synthetic Likelihood
- Efficient acquisition rules for model-based approximate Bayesian computation
- Gaussian process modelling in approximate Bayesian computation to estimate horizontal gene transfer in bacteria
- Gaussian processes for machine learning.
- scientific article; zbMATH DE number 3583145 (Why is no real title available?)
- scientific article; zbMATH DE number 2117879 (Why is no real title available?)
- scientific article; zbMATH DE number 6433488 (Why is no real title available?)
- scientific article; zbMATH DE number 6276166 (Why is no real title available?)
- Laplace approximation for logistic Gaussian process density estimation and regression
- Learning with submodular functions: a convex optimization perspective
- Likelihood-free inference via classification
- Median loss decision theory
- Near-optimal sensor placements in Gaussian processes: theory, efficient algorithms and empirical studies
- Nonparametric guidance of autoencoder representations using label information
- Posterior consistency for Gaussian process approximations of Bayesian posterior distributions
- Probabilistic integration: a role in statistical computation?
- Probabilistic numerics and uncertainty in computations
- Robust Bayesian synthetic likelihood via a semi-parametric approach
- Sequential design of computer experiments for the estimation of a probability of failure
- Sequential design of computer experiments for the solution of Bayesian inverse problems
- Stochastic kriging for simulation metamodeling
- The Bayesian Choice
Cited in
(16)- Evaluating Gaussian process metamodels and sequential designs for noisy level set estimation
- Parallel partial Gaussian process emulation for computer models with massive output
- scientific article; zbMATH DE number 7625191 (Why is no real title available?)
- scientific article; zbMATH DE number 7164769 (Why is no real title available?)
- Sequentially guided MCMC proposals for synthetic likelihoods and correlated synthetic likelihoods
- Residual-based error correction for neural operator accelerated Infinite-dimensional Bayesian inverse problems
- Bayesian estimation of large-scale simulation models with Gaussian process regression surrogates
- An efficient two-stage algorithm for parameter identification of non-linear state-space models-based on Gaussian process regression
- Fast post-process Bayesian inference with variational sparse Bayesian quadrature
- Warped gradient-enhanced Gaussian process surrogate models for exponential family likelihoods with intractable normalizing constants
- A survey of Monte Carlo methods for noisy and costly densities with application to reinforcement learning and ABC
- Introduction to Gaussian process regression in Bayesian inverse problems, with new results on experimental design for weighted error measures
- Likelihood-free frequentist inference: bridging classical statistics and machine learning for reliable simulator-based inference
- Batch Sequential Experimental Design for Calibration of Stochastic Simulation Models
- Estimating intractable posterior distributions through Gaussian process regression and metropolis-adjusted Langevin procedure
- Parallel cross-validation: a scalable fitting method for Gaussian process models
This page was built for publication: Parallel Gaussian process surrogate Bayesian inference with noisy likelihood evaluations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2057377)
