On the convergence of the Laplace approximation and noise-level-robustness of Laplace-based Monte Carlo methods for Bayesian inverse problems
From MaRDI portal
Publication:2194045
Convergence of probability measures (60B10) Bayesian inference (62F15) Monte Carlo methods (65C05) Numerical optimization and variational techniques (65K10) Bayesian problems; characterization of Bayes procedures (62C10) Random measures (60G57) Stability and convergence of numerical methods for initial value and initial-boundary value problems involving PDEs (65M12) Error bounds for initial value and initial-boundary value problems involving PDEs (65M15) Numerical methods for inverse problems for initial value and initial-boundary value problems involving PDEs (65M32)
Abstract: The Bayesian approach to inverse problems provides a rigorous framework for the incorporation and quantification of uncertainties in measurements, parameters and models. We are interested in designing numerical methods which are robust w.r.t. the size of the observational noise, i.e., methods which behave well in case of concentrated posterior measures. The concentration of the posterior is a highly desirable situation in practice, since it relates to informative or large data. However, it can pose a computational challenge for numerical methods based on the prior or reference measure. We propose to employ the Laplace approximation of the posterior as the base measure for numerical integration in this context. The Laplace approximation is a Gaussian measure centered at the maximum a-posteriori estimate and with covariance matrix depending on the logposterior density. We discuss convergence results of the Laplace approximation in terms of the Hellinger distance and analyze the efficiency of Monte Carlo methods based on it. In particular, we show that Laplace-based importance sampling and Laplace-based quasi-Monte-Carlo methods are robust w.r.t. the concentration of the posterior for large classes of posterior distributions and integrands whereas prior-based importance sampling and plain quasi-Monte Carlo are not. Numerical experiments are presented to illustrate the theoretical findings.
Recommendations
- Non-asymptotic error estimates for the Laplace approximation in Bayesian inverse problems
- Convergence rate for the Bayesian approach to linear inverse problems
- A Bayesian approach to multiscale inverse problems using the sequential Monte Carlo method
- Quasi-Monte Carlo Bayesian estimation under Besov priors in elliptic inverse problems
- Convergence analysis of surrogate-based methods for Bayesian inverse problems
- Approximation of Bayesian Inverse Problems for PDEs
- On the well-posedness of Bayesian inverse problems
- Sequential Monte Carlo methods for Bayesian elliptic inverse problems
- Bayesian inverse problems with \(l_1\) priors: a randomize-then-optimize approach
- A hierarchical Bayesian setting for an inverse problem in linear parabolic PDEs with noisy boundary conditions
Cites work
- scientific article; zbMATH DE number 2221907 (Why is no real title available?)
- A Fast and Scalable Method for A-Optimal Design of Experiments for Infinite-dimensional Bayesian Nonlinear Inverse Problems
- A Hierarchical Multilevel Markov Chain Monte Carlo Algorithm with Applications to Uncertainty Quantification in Subsurface Flow
- A Review of Modern Computational Algorithms for Bayesian Optimal Design
- A stochastic collocation approach to Bayesian inference in inverse problems
- An analysis of Bayesian inference for nonparametric regression
- Application of quasi-Monte Carlo methods to elliptic PDEs with random diffusion coefficients: a survey of analysis and implementation
- Asymptotic Statistics
- Asymptotic approximations of integrals
- Bernstein-von Mises theorems for statistical inverse problems. I: Schrödinger equation
- Combinatorics of partial derivatives
- Complexity analysis of accelerated MCMC methods for Bayesian inversion
- Convergence rates of posterior distributions.
- Dimension-Independent MCMC Sampling for Inverse Problems with Non-Gaussian Priors
- Fast Bayesian experimental design: Laplace-based importance sampling for the expected information gain
- Fast estimation of expected information gains for Bayesian experimental designs based on Laplace approximations
- Frequentist coverage of adaptive nonparametric Bayesian credible sets
- Gaussian approximations for probability measures on \(\mathbb R^d\)
- Hessian-based adaptive sparse quadrature for infinite-dimensional Bayesian inverse problems
- High-dimensional integration: The quasi-Monte Carlo way
- Higher order quasi-Monte Carlo integration for holomorphic, parametric operator equations
- Inverse problems: a Bayesian perspective
- Kullback-Leibler approximation for probability measures on infinite dimensional spaces
- Multilevel higher-order quasi-Monte Carlo Bayesian estimation
- Nonparametric Bernstein-von Mises theorems in Gaussian white noise
- On Choosing and Bounding Probability Metrics
- On a generalization of the preconditioned Crank-Nicolson metropolis algorithm
- On the Bernstein-von Mises approximation of posterior distributions
- On the Bernstein-von Mises phenomenon for nonparametric Bayes procedures
- On the Bernstein-von Mises theorem with infinite-dimensional parameters
- On the consistency of Bayes estimates
- Quasi-Monte Carlo and multilevel Monte Carlo methods for computing posterior expectations in elliptic inverse problems
- Scaling limits in computational Bayesian inversion
- Sparse, adaptive Smolyak quadratures for Bayesian inverse problems
- Sparsity in Bayesian inversion of parametric operator equations
- Statistical and computational inverse problems.
- The Bernstein-von Mises theorem under misspecification
Cited in
(31)- Residual-based error correction for neural operator accelerated Infinite-dimensional Bayesian inverse problems
- Isogeometric multilevel quadrature for forward and inverse random acoustic scattering
- Multilevel Monte Carlo estimation of the expected value of sample information
- A Variational Inference Approach to Inverse Problems with Gamma Hyperpriors
- Wasserstein convergence rates of increasingly concentrating probability measures
- Learning physics-based models from data: perspectives from inverse problems and model reduction
- Non-asymptotic error estimates for the Laplace approximation in Bayesian inverse problems
- Stability of Gibbs posteriors from the Wasserstein loss for Bayesian full waveform inversion
- Re-thinking high-dimensional mathematical statistics. Abstracts from the workshop held May 15--21, 2022
- Scaling limits in computational Bayesian inversion
- On log-concave approximations of high-dimensional posterior measures and stability properties in non-linear inverse problems
- Multimodal information gain in Bayesian design of experiments
- Variational inference for nonlinear inverse problems via neural net kernels: comparison to Bayesian neural networks, application to topology optimization
- Dimension Free Nonasymptotic Bounds on the Accuracy of High-Dimensional Laplace Approximation
- Certified dimension reduction in nonlinear Bayesian inverse problems
- Quasi-Monte Carlo and multilevel Monte Carlo methods for computing posterior expectations in elliptic inverse problems
- Context-Aware Surrogate Modeling for Balancing Approximation and Sampling Costs in Multifidelity Importance Sampling and Bayesian Inverse Problems
- The computational asymptotics of Gaussian variational inference and the Laplace approximation
- Horseshoe Priors for Edge-Preserving Linear Bayesian Inversion
- Optimal experimental design: formulations and computations
- On the Error Rate of Importance Sampling with Randomized Quasi-Monte Carlo
- Shape reconstructions by using plasmon resonances
- Low-rank tensor reconstruction of concentrated densities with application to Bayesian inversion
- Determination of the reaction coefficient in a time dependent nonlocal diffusion process
- Sparse approximation of triangular transports. I: The finite-dimensional case
- Small-noise approximation for Bayesian optimal experimental design with nuisance uncertainty
- Bayesian neural network priors for edge-preserving inversion
- Unbiased MLMC stochastic gradient-based optimization of Bayesian experimental designs
- A hybrid Gibbs sampler for edge-preserving tomographic reconstruction with uncertain view angles
- Shape reconstructions by using plasmon resonances with enhanced sensitivity
- Adaptive operator learning for infinite-dimensional Bayesian inverse problems
This page was built for publication: On the convergence of the Laplace approximation and noise-level-robustness of Laplace-based Monte Carlo methods for Bayesian inverse problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2194045)
