Parallel local approximation MCMC for expensive models
From MaRDI portal
Abstract: Performing Bayesian inference via Markov chain Monte Carlo (MCMC) can be exceedingly expensive when posterior evaluations invoke the evaluation of a computationally expensive model, such as a system of partial differential equations. In recent work [Conrad et al. JASA 2016, arXiv:1402.1694], we described a framework for constructing and refining local approximations of such models during an MCMC simulation. These posterior--adapted approximations harness regularity of the model to reduce the computational cost of inference while preserving asymptotic exactness of the Markov chain. Here we describe two extensions of that work. First, we prove that samplers running in parallel can collaboratively construct a shared posterior approximation while ensuring ergodicity of each associated chain, providing a novel opportunity for exploiting parallel computation in MCMC. Second, focusing on the Metropolis--adjusted Langevin algorithm, we describe how a proposal distribution can successfully employ gradients and other relevant information extracted from the approximation. We investigate the practical performance of our strategies using two challenging inference problems, the first in subsurface hydrology and the second in glaciology. Using local approximations constructed via parallel chains, we successfully reduce the run time needed to characterize the posterior distributions in these problems from days to hours and from months to days, respectively, dramatically improving the tractability of Bayesian inference.
Recommendations
- Rate-optimal refinement strategies for local approximation MCMC
- Parallel MCMC with generalized elliptical slice sampling
- Localization for MCMC: sampling high-dimensional posterior distributions with local structure
- Parallel algorithms for Bayesian inference in spatial Gaussian models
- Bayesian inversion by parallel interacting Markov chains
Cites work
- A central limit theorem for convex sets
- A class of Wasserstein metrics for probability distributions
- A computational framework for infinite-dimensional Bayesian inverse problems. II: stochastic Newton MCMC with application to ice sheet flow inverse problems
- A stochastic collocation approach to Bayesian inference in inverse problems
- A stochastic Newton MCMC method for large-scale statistical inverse problems with application to seismic inversion
- Adaptive construction of surrogates for the Bayesian solution of inverse problems
- An adaptive Metropolis algorithm
- Approximation of Bayesian Inverse Problems for PDEs
- Bayesian calibration of computer models. (With discussion)
- Bayesian computation: a summary of the current state, and samples backwards and forwards
- Design and analysis of computer experiments. With comments and a rejoinder by the authors
- Exact sampling with coupled Markov chains and applications to statistical mechanics
- scientific article; zbMATH DE number 1843488 (Why is no real title available?)
- Introduction to Derivative-Free Optimization
- Learn from thy neighbor: parallel-chain and regional adaptive MCMC
- Locally contractive iterated function systems
- Markov Chain Monte Carlo Convergence Diagnostics: A Comparative Review
- Monte Carlo errors with less errors
- One-shot coupling for certain stochastic recursive sequences.
- Optimal Scaling of Discrete Approximations to Langevin Diffusions
- Optimal Transport
- Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
- Scalable posterior approximations for large-scale Bayesian inverse problems via likelihood-informed parameter and state reduction
- Sparse-grid, reduced-basis Bayesian inversion
- Statistical and computational inverse problems.
- Stochastic spectral methods for efficient Bayesian solution of inverse problems
- The design and analysis of computer experiments.
- The Geometry of Random Fields
- Uncertainty Quantification and Weak Approximation of an Elliptic Inverse Problem
Cited in
(26)- Powered embarrassing parallel MCMC sampling in Bayesian inference, a weighted average intuition
- SIMD parallel MCMC sampling with applications for big-data Bayesian analytics
- X-TMCMC: adaptive kriging for Bayesian inverse modeling
- Accelerating MCMC via Kriging-based adaptive independent proposals and delayed rejection
- Anytime parallel tempering
- Emulation-accelerated Hamiltonian Monte Carlo algorithms for parameter estimation and uncertainty quantification in differential equation models
- Bayesian model inversion using stochastic spectral embedding
- Variational inference with NoFAS: normalizing flow with adaptive surrogate for computationally expensive models
- Rate-optimal refinement strategies for local approximation MCMC
- Adaptive method for indirect identification of the statistical properties of random fields in a Bayesian framework
- Localization for MCMC: sampling high-dimensional posterior distributions with local structure
- Adaptive multi-fidelity polynomial chaos approach to Bayesian inference in inverse problems
- A transport-based multifidelity preconditioner for Markov chain Monte Carlo
- Parallel partial Gaussian process emulation for computer models with massive output
- Efficient MCMC for climate model parameter estimation: parallel adaptive chains and early rejection
- Parallel MCMC with generalized elliptical slice sampling
- Sampling and Statistical Physics via Symmetry
- Scalable Optimization-Based Sampling on Function Space
- Adaptive sequential posterior simulators for massively parallel computing environments
- A Bayesian Approach to Modeling Biological Pattern Formation with Limited Data
- A reduced basis ensemble Kalman method
- Multilevel Delayed Acceptance MCMC
- A MCMC method based on surrogate model and Gaussian process parameterization for infinite Bayesian PDE inversion
- Deterministic Sampling of Expensive Posteriors Using Minimum Energy Designs
- Adaptive reduced tempering for Bayesian inverse problems and rare event simulation
- A survey of Monte Carlo methods for noisy and costly densities with application to reinforcement learning and ABC
This page was built for publication: Parallel local approximation MCMC for expensive models
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4636376)
