Multilevel higher-order quasi-Monte Carlo Bayesian estimation

DOI10.1142/S021820251750021XMaRDI QIDQ2988720zbMATH OpenOpenAlexFDO

Authors Josef Dick, Robert Nicholas Gantner, Christoph Schwab, Q. T. Le Gia

Publication date 19 May 2017

Published in M\(^3\)AS. Mathematical Models \& Methods in Applied Sciences (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1611.08324

zbMATH Keywords

convergenceuncertainty quantificationalgorithmerror analysisPetrov-Galerkin methodnumerical testsBayesian inverse problemsmultilevel approximationshigher-order quasi-Monte Carlo quadraturenon-linear parametric operator equations

Mathematics Subject Classification ID

Bayesian inference (62F15) Monte Carlo methods (65C05) Probabilistic methods, particle methods, etc. for boundary value problems involving PDEs (65N75) Numerical methods for inverse problems for boundary value problems involving PDEs (65N21) Stochastic particle methods (65C35) Stability and convergence of numerical methods for boundary value problems involving PDEs (65N12) Error bounds for boundary value problems involving PDEs (65N15)

Abstract: We propose and analyze deterministic multilevel approximations for Bayesian inversion of operator equations with uncertain distributed parameters, subject to additive Gaussian measurement data. The algorithms use a multilevel (ML) approach based on deterministic, higher order quasi-Monte Carlo (HoQMC) quadrature for approximating the high-dimensional expectations, which arise in the Bayesian estimators, and a Petrov-Galerkin (PG) method for approximating the solution to the underlying partial differential equation (PDE). This extends the previous single-level approach from [J. Dick, R. N. Gantner, Q. T. Le Gia and Ch. Schwab, Higher order Quasi-Monte Carlo integration for Bayesian Estimation. Report 2016-13, Seminar for Applied Mathematics, ETH Z"urich (in review)]. We obtain sufficient conditions which allow us to achieve arbitrarily high, algebraic convergence rates in terms of work, which are independent of the dimension of the parameter space. The convergence rates are limited only by the spatial regularity of the forward problem,the discretization order achieved by the Petrov Galerkin discretization, and by the sparsity of the uncertainty parametrization. We provide detailed numerical experiments for linear elliptic problems in two space dimensions, with $s=1024$ parameters characterizing the uncertain input, confirming the theory and showing that the ML HoQMC algorithms outperform, in terms of error vs.~computational work, both multilevel Monte Carlo (MLMC) methods and single-level (SL) HoQMC methods.

Recommendations

• Multilevel Monte Carlo in approximate Bayesian computation
• scientific article; zbMATH DE number 7626793
• Multilevel Markov Chain Monte Carlo
• scientific article; zbMATH DE number 2000348
• Multilevel sequential Monte Carlo samplers
• Multilevel sequential Monte Carlo for Bayesian inverse problems
• High-dimensional and higher-order multifidelity Monte Carlo estimators
• Continuous Level Monte Carlo and Sample-Adaptive Model Hierarchies

Cites work

• A Hierarchical Multilevel Markov Chain Monte Carlo Algorithm with Applications to Uncertainty Quantification in Subsurface Flow
• Bounds on Walsh coefficients by dyadic difference and a new Koksma-Hlawka type inequality for quasi-Monte Carlo integration
• Breaking the curse of dimensionality in sparse polynomial approximation of parametric PDEs
• Complexity analysis of accelerated MCMC methods for Bayesian inversion
• Computational higher order quasi-Monte Carlo integration
• Consistency, stability, a priori and a posteriori errors for Petrov- Galerkin methods applied to nonlinear problems
• Dimension truncation in QMC for affine-parametric operator equations
• Finite dimensional approximation of nonlinear problems. I: Branches of nonsingular solutions
• Finite element approximation of the Navier-Stokes equations
• Higher order QMC Petrov-Galerkin discretization for affine parametric operator equations with random field inputs
• Higher order quasi-Monte Carlo integration for holomorphic, parametric operator equations
• Inverse problems: a Bayesian perspective
• Karhunen-Loève approximation of random fields by generalized fast multipole methods
• Multi-level quasi-Monte Carlo finite element methods for a class of elliptic PDEs with random coefficients
• Multilevel Quasi-Monte Carlo methods for lognormal diffusion problems
• Multilevel accelerated quadrature for PDEs with log-normally distributed diffusion coefficient
• Multilevel higher order QMC Petrov-Galerkin discretization for affine parametric operator equations
• Multilevel higher-order quasi-Monte Carlo Bayesian estimation
• Multilevel sequential Monte Carlo samplers
• Quasi-Monte Carlo and multilevel Monte Carlo methods for computing posterior expectations in elliptic inverse problems
• Quasi-Monte Carlo finite element methods for a class of elliptic partial differential equations with random coefficients
• Scaling limits in computational Bayesian inversion
• Sparse adaptive approximation of high dimensional parametric initial value problems
• Sparse tensor discretizations of high-dimensional parametric and stochastic PDEs
• Sparse, adaptive Smolyak quadratures for Bayesian inverse problems
• Sparsity in Bayesian inversion of parametric operator equations
• Walsh Spaces Containing Smooth Functions and Quasi–Monte Carlo Rules of Arbitrary High Order

Cited in

(34)

• Deep learning in high dimension: neural network expression rates for generalized polynomial chaos expansions in UQ
• Quasi-Monte Carlo and multilevel Monte Carlo methods for computing posterior expectations in elliptic inverse problems
• A weighted discrepancy bound of quasi-Monte Carlo importance sampling
• Constructive deep ReLU neural network approximation
• Convergence rates for a class of estimators based on Stein's method
• Multifidelity multilevel Monte Carlo to accelerate approximate Bayesian parameter inference for partially observed stochastic processes
• Large deformation shape uncertainty quantification in acoustic scattering
• Higher-order quasi-Monte Carlo for Bayesian shape inversion
• Electromagnetic wave scattering by random surfaces: shape holomorphy
• Improved Efficiency of a Multi-Index FEM for Computational Uncertainty Quantification
• Quasi-Monte Carlo Bayesian estimation under Besov priors in elliptic inverse problems
• Bayesian model inversion using stochastic spectral embedding
• Analysis of a multilevel Markov chain Monte Carlo finite element method for Bayesian inversion of log-normal diffusions
• Wavenumber-Explicit Parametric Holomorphy of Helmholtz Solutions in the Context of Uncertainty Quantification
• Deep neural network expression of posterior expectations in Bayesian PDE inversion
• Extrapolated polynomial lattice rule integration in computational uncertainty quantification
• Multilevel approximation of parametric and stochastic PDES
• Parametric shape holomorphy of boundary integral operators with applications
• Higher order quasi-Monte Carlo integration for Bayesian PDE inversion
• Scaling limits in computational Bayesian inversion
• Domain uncertainty quantification in computational electromagnetics
• Convergence of adaptive stochastic collocation with finite elements
• Multilevel higher-order quasi-Monte Carlo Bayesian estimation
• Multilevel Markov Chain Monte Carlo
• Isogeometric multilevel quadrature for forward and inverse random acoustic scattering
• On the convergence of the Laplace approximation and noise-level-robustness of Laplace-based Monte Carlo methods for Bayesian inverse problems
• Higher-order quasi-Monte Carlo training of deep neural networks
• Sparse Compression of Expected Solution Operators
• Multilevel adaptive sparse Leja approximations for Bayesian inverse problems
• Multilevel Monte Carlo in approximate Bayesian computation
• Multilevel Monte Carlo estimation of the expected value of sample information
• Multilevel Markov chain Monte Carlo for Bayesian inversion of parabolic partial differential equations under Gaussian prior
• Multilevel quadrature for elliptic parametric partial differential equations in case of polygonal approximations of curved domains
• Multilevel Quasi-Monte Carlo Uncertainty Quantification for Advection-Diffusion-Reaction

This page was built for publication: Multilevel higher-order quasi-Monte Carlo Bayesian estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2988720)