Discretization-invariant Bayesian inversion and Besov space priors

DOI10.3934/IPI.2009.3.87zbMATH Open1191.62046arXiv0901.4220OpenAlexW2963777151WikidataQ109553804 ScholiaQ109553804MaRDI QIDQ2268280FDOQ2268280

S. Siltanen, Eero Saksman, M. Lassas

Publication date: 10 March 2010

Published in: Inverse Problems and Imaging (Search for Journal in Brave)

Abstract: Bayesian solution of an inverse problem for indirect measurement $M=AU+mathcalE$ is considered, where $U$ is a function on a domain of $R^d$. Here $A$ is a smoothing linear operator and $mathcalE$ is Gaussian white noise. The data is a realization $m_k$ of the random variable $M_k=P_{k}AU+P_{k}mathcalE$, where $P_k$ is a linear, finite dimensional operator related to measurement device. To allow computerized inversion, the unknown is discretized as $U_n=T_{n}U$, where $T_n$ is a finite dimensional projection, leading to the computational measurement model $M_{kn}=P_{k}A{U}_n+P_{k}mathcalE$. Bayes formula gives then the posterior distribution $p{i}_{kn}(u_n|m_{kn})simp{i}_n(u_n)exp(-1/2|m_{kn}-P_{k}A{u}_n{|}_2^2)$ in $R^d$, and the mean $U_{kn}^{CM}:=int{u}_{n}p{i}_{kn}(u_n|m_k)d{u}_n$ is considered as the reconstruction of $U$. We discuss a systematic way of choosing prior distributions $prio{r}_n$ for all $ngeq{n}_0>0$ by achieving them as projections of a distribution in a infinite-dimensional limit case. Such choice of prior distributions is {em discretization-invariant} in the sense that $prio{r}_n$ represent the same {em a priori} information for all $n$ and that the mean $U_{kn}^{CM}$ converges to a limit estimate as $k,noinfty$. Gaussian smoothness priors and wavelet-based Besov space priors are shown to be discretization invariant. In particular, Bayesian inversion in dimension two with $B_{11}^1$ prior is related to penalizing the $el{l}^1$ norm of the wavelet coefficients of $U$.

Full work available at URL: https://arxiv.org/abs/0901.4220

zbMATH Keywords

inverse problemreconstructionBesov spacewaveletBayesian inversionstatistical inversiondiscretization invariance

Mathematics Subject Classification ID

Bayesian inference (62F15) Nontrigonometric harmonic analysis involving wavelets and other special systems (42C40) Applications of functional analysis in probability theory and statistics (46N30) Probabilistic models, generic numerical methods in probability and statistics (65C20) Functional limit theorems; invariance principles (60F17)

Cited In (73)

• A Bayesian approach to wavelet-based modelling of discontinuous functions applied to inverse problems
• Adaptive inference over Besov spaces in the white noise model using \(p\)-exponential priors
• A Bayesian approach for consistent reconstruction of inclusions
• Certified coordinate selection for high-dimensional Bayesian inversion with Laplace prior
• Scalable conditional deep inverse Rosenblatt transports using tensor trains and gradient-based dimension reduction
• Edge-Promoting Adaptive Bayesian Experimental Design for X-ray Imaging
• Reconciling Bayesian and Perimeter Regularization for Binary Inversion
• Multilevel Monte Carlo FEM for elliptic PDEs with Besov random tree priors
• Are Minimizers of the Onsager–Machlup Functional Strong Posterior Modes?
• Besov-Laplace priors in density estimation: optimal posterior contraction rates and adaptation
• Heavy-tailed Bayesian nonparametric adaptation
• Field Dynamics Inference for Local and Causal Interactions
• Bayesian inversion with Student's \(t\) priors based on Gaussian scale mixtures
• An Order-Theoretic Perspective on Modes and Maximum A Posteriori Estimation in Bayesian Inverse Problems
• Laplace priors and spatial inhomogeneity in Bayesian inverse problems
• Spectral neural operators
• Frame-constrained total variation regularization for white noise regression
• Sparsity-promoting and edge-preserving maximum a posteriori estimators in non-parametric Bayesian inverse problems
• Oracle-type posterior contraction rates in Bayesian inverse problems
• Nonparametric Bayesian inference for reversible multidimensional diffusions
• On the lifting of deterministic convergence rates for inverse problems with stochastic noise
• Wavelet-based priors accelerate maximum-a-posteriori optimization in Bayesian inverse problems
• Bayesian approach for inverse interior scattering problems with limited aperture
• An approximate empirical Bayesian method for large-scale linear-Gaussian inverse problems
• Hyperpriors for Matérn fields with applications in Bayesian inversion
• Bernstein-von Mises theorems for statistical inverse problems. I: Schrödinger equation
• Optimal experimental design for infinite-dimensional Bayesian inverse problems governed by PDEs: a review
• Γ-convergence of Onsager–Machlup functionals: II. Infinite product measures on Banach spaces
• Well-posed Bayesian inverse problems and heavy-tailed stable quasi-Banach space priors
• Shearlet-based regularization in statistical inverse learning with an application to x-ray tomography
• Quasi-Monte Carlo Bayesian estimation under Besov priors in elliptic inverse problems
• Cauchy Markov random field priors for Bayesian inversion
• Regularization with metric double integrals of functions with values in a set of vectors
• On unifying randomized methods for inverse problems
• Diffusion tensor imaging with deterministic error bounds
• Deep neural network expression of posterior expectations in Bayesian PDE inversion
• Cauchy difference priors for edge-preserving Bayesian inversion
• A Fast and Scalable Method for A-Optimal Design of Experiments for Infinite-dimensional Bayesian Nonlinear Inverse Problems
• Bernstein--von Mises Theorems and Uncertainty Quantification for Linear Inverse Problems
• Spectral gaps for a Metropolis-Hastings algorithm in infinite dimensions
• Data-free likelihood-informed dimension reduction of Bayesian inverse problems
• Generalized Modes in Bayesian Inverse Problems
• Bayesian Inverse Problems with $l_1$ Priors: A Randomize-Then-Optimize Approach
• Two Metropolis--Hastings Algorithms for Posterior Measures with Non-Gaussian Priors in Infinite Dimensions
• Material-separating regularizer for multi-energy x-ray tomography
• Data-driven forward discretizations for Bayesian inversion
• A Bayesian-based approach to improving acoustic Born waveform inversion of seismic data for viscoelastic media
• Multiresolution Parameter Choice Method for Total Variation Regularized Tomography
• The influence of numerical error on parameter estimation and uncertainty quantification for advective PDE models
• Mathematics of Electron Tomography
• The Bayesian formulation of EIT: analysis and algorithms
• Regularized posteriors in linear ill-posed inverse problems
• Bayesian Probabilistic Numerical Methods
• Well-Posed Bayesian Inverse Problems: Priors with Exponential Tails
• Prior normalization for certified likelihood-informed subspace detection of Bayesian inverse problems
• Bayesian neural network priors for edge-preserving inversion
• Joint estimation of Robin coefficient and domain boundary for the Poisson problem
• Bayesian inversion with α-stable priors
• Bayesian inversion techniques for stochastic partial differential equations
• Horseshoe Priors for Edge-Preserving Linear Bayesian Inversion
• A Hybrid Gibbs Sampler for Edge-Preserving Tomographic Reconstruction with Uncertain View Angles
• Γ -convergence of Onsager–Machlup functionals: I. With applications to maximum a posteriori estimation in Bayesian inverse problems
• Infinite-dimensional inverse problems with finite measurements
• Rates of contraction of posterior distributions based on \(p\)-exponential priors
• A scalable algorithm for MAP estimators in Bayesian inverse problems with Besov priors
• Modern regularization methods for inverse problems
• Uncertainty Quantification of Inclusion Boundaries in the Context of X-Ray Tomography
• Optimal scaling and diffusion limits for the Langevin algorithm in high dimensions
• Solving inverse problems using data-driven models
• Random tree Besov priors -- towards fractal imaging
• Can one use total variation prior for edge-preserving Bayesian inversion?
• Whittle-Matérn priors for Bayesian statistical inversion with applications in electrical impedance tomography
• Bayesian inversion for electrical-impedance tomography in medical imaging using the nonlinear Poisson-Boltzmann equation

This page was built for publication: Discretization-invariant Bayesian inversion and Besov space priors

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2268280)