Accelerating Asymptotically Exact MCMC for Computationally Intensive Models via Local Approximations

Abstract

We construct a new framework for accelerating Markov chain Monte Carlo in posterior sampling problems where standard methods are limited by the computational cost of the likelihood, or of numerical models embedded therein. Our approach introduces local approximations of these models into the Metropolis-Hastings kernel, borrowing ideas from deterministic approximation theory, optimization, and experimental design. Previous efforts at integrating approximate models into inference typically sacrifice either the sampler's exactness or efficiency; our work seeks to address these limitations by exploiting useful convergence characteristics of local approximations. We prove the ergodicity of our approximate Markov chain, showing that it samples asymptotically from the \emph{exact} posterior distribution of interest. We describe variations of the algorithm that employ either local polynomial approximations or local Gaussian process regressors. Our theoretical results reinforce the key observation underlying this paper: when the likelihood has some \emph{local} regularity, the number of model evaluations per MCMC step can be greatly reduced without biasing the Monte Carlo average. Numerical experiments demonstrate multiple order-of-magnitude reductions in the number of forward model evaluations used in representative ODE and PDE inference problems, with both synthetic and real data.

A major update of the theory and examples

References in corpus (4)

• The pseudo-marginal approach for efficient Monte Carlo computations
• Sparse Pseudospectral Approximation Method
• Adaptive construction of surrogates for the Bayesian solution of inverse problems
• Convergence of adaptive and interacting Markov chain Monte Carlo algorithms

Cited by in corpus (22)

• Calibrate, Emulate, Sample
• Geometric MCMC for Infinite-Dimensional Inverse Problems
• Adaptive multi-fidelity polynomial chaos approach to Bayesian inference in inverse problems
• Polynomial chaos expansions for dependent random variables
• Surrogate and inverse modeling for two-phase flow in porous media via theory-guided convolutional neural network
• Surrogate-Based Bayesian Inverse Modeling of the Hydrological System: An Adaptive Approach Considering Surrogate Approximation Error
• Deterministic Sampling of Expensive Posteriors Using Minimum Energy Designs
• Adaptive Model Refinement Approach for Bayesian Uncertainty Quantification in Turbulence Model
• Adaptive numerical designs for the calibration of computer codes
• Bayesian model inversion using stochastic spectral embedding
• Adaptive quadrature schemes for Bayesian inference via active learning
• Stein variational gradient descent with local approximations
• Variational Inference with NoFAS: Normalizing Flow with Adaptive Surrogate for Computationally Expensive Models
• Goal-oriented adaptive surrogate construction for stochastic inversion
• Towards fast machine-learning-assisted Bayesian posterior inference of microseismic event location and source mechanism
• A survey of Monte Carlo methods for noisy and costly densities with application to reinforcement learning and ABC
• Goal-Oriented A-Posteriori Estimation of Model Error as an Aid to Parameter Estimation
• Accelerating sequential Monte Carlo with surrogate likelihoods
• Iterative Construction of Gaussian Process Surrogate Models for Bayesian Inference
• Deep Importance Sampling based on Regression for Model Inversion and Emulation
• Approximate Gibbs Sampler for Efficient Inference of Hierarchical Bayesian Models for Grouped Count Data
• FLARE MCMC: Fidelity-based Layer-Adaptive REcursive proposals for MCMC