Generalized Hybrid Iterative Methods for Large-Scale Bayesian Inverse Problems
From MaRDI portal
Abstract: We develop a generalized hybrid iterative approach for computing solutions to large-scale Bayesian inverse problems. We consider a hybrid algorithm based on the generalized Golub-Kahan bidiagonalization for computing Tikhonov regularized solutions to problems where explicit computation of the square root and inverse of the covariance kernel for the prior covariance matrix is not feasible. This is useful for large-scale problems where covariance kernels are defined on irregular grids or are only available via matrix-vector multiplication, e.g., those from the Mat'{e}rn class. We show that iterates are equivalent to LSQR iterates applied to a directly regularized Tikhonov problem, after a transformation of variables, and we provide connections to a generalized singular value decomposition filtered solution. Our approach shares many benefits of standard hybrid methods such as avoiding semi-convergence and automatically estimating the regularization parameter. Numerical examples from image processing demonstrate the effectiveness of the described approaches.
Recommendations
- Efficient derivative-free Bayesian inference for large-scale inverse problems
- Generalized Modes in Bayesian Inverse Problems
- An approximate empirical Bayesian method for large-scale linear-Gaussian inverse problems
- A Bayesian Numerical Homogenization Method for Elliptic Multiscale Inverse Problems
- Sparsity Promoting Hybrid Solvers for Hierarchical Bayesian Inverse Problems
- Hybrid projection methods for large-scale inverse problems with mixed Gaussian priors
- Fast Bayesian inversion for high dimensional inverse problems
- Approximation of Bayesian Inverse Problems for PDEs
- Iterative Bayesian inversion with Gaussian mixtures: finite sample implementation and large sample asymptotics
- Stochastic spectral methods for efficient Bayesian solution of inverse problems
Cites work
- A Bidiagonalization-Regularization Procedure for Large Scale Discretizations of Ill-Posed Problems
- A Block Lanczos Method for Computing the Singular Values and Corresponding Singular Vectors of a Matrix
- A Golub-Kahan-type reduction method for matrix pairs
- A hybrid LSMR algorithm for large-scale Tikhonov regularization
- A Projection‐Based Approach to General‐Form Tikhonov Regularization
- A weighted-GCV method for Lanczos-hybrid regularization
- An efficient iterative approach for large-scale separable nonlinear inverse problems
- An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach
- An iterative method for Tikhonov regularization with a general linear regularization operator
- Calculating the Singular Values and Pseudo-Inverse of a Matrix
- Choosing regularization parameters in iterative methods for ill-posed problems
- Computational Methods for Inverse Problems
- De-noising by soft-thresholding
- Deblurring Images
- Discrete inverse problems. Insight and algorithms.
- Efficient computation of linearized cross-covariance and auto-covariance matrices of interdependent quantities
- Gaussian processes for machine learning.
- Generalized Arnoldi-Tikhonov method for sparse reconstruction
- Generalized Cross-Validation as a Method for Choosing a Good Ridge Parameter
- Generalized Golub-Kahan bidiagonalization and stopping criteria
- Generalizing the Singular Value Decomposition
- GKB-FP: An algorithm for large-scale discrete ill-posed problems
- scientific article; zbMATH DE number 1953444 (Why is no real title available?)
- scientific article; zbMATH DE number 2021122 (Why is no real title available?)
- scientific article; zbMATH DE number 1460605 (Why is no real title available?)
- scientific article; zbMATH DE number 3408799 (Why is no real title available?)
- Hybrid and iteratively reweighted regularization by unbiased predictive risk and weighted GCV for projected systems
- Introduction to Bayesian Scientific Computing
- Iterated preconditioned LSQR method for inverse problems on unstructured grids
- Iterative regularization with minimum-residual methods
- Large-scale stochastic linear inversion using hierarchical matrices. Illustrated with an application to crosswell tomography in seismic imaging
- LSMR: An Iterative Algorithm for Sparse Least-Squares Problems
- LSQR: An Algorithm for Sparse Linear Equations and Sparse Least Squares
- Numerical methods for coupled super-resolution
- On Krylov projection methods and Tikhonov regularization
- On the Rates of Convergence of the Lanczos and the Block-Lanczos Methods
- Preconditioned iterative methods for linear discrete ill-posed problems from a Bayesian inversion perspective
- Priorconditioners for linear systems
- Rank-one modification of the symmetric eigenproblem
- Regularization methods for large-scale problems
- Regularization parameter estimation for large-scale Tikhonov regularization using a priori information
- Regularization tools: A Matlab package for analysis and solution of discrete ill-posed problems
- Some Modified Matrix Eigenvalue Problems
- The behavior of eigenvalues and singular values under perturbations of restricted rank
- The regularizing effect of the Golub-Kahan iterative bidiagonalization and revealing the noise level in the data
- Tikhonov regularization based on generalized Krylov subspace methods
Cited in
(31)- A residual-driven adaptive Gaussian mixture approximation for Bayesian inverse problems
- Kryging: geostatistical analysis of large-scale datasets using Krylov subspace methods
- Weighted tensor Golub-Kahan-Tikhonov-type methods applied to image processing using a t-product
- A joint bidiagonalization based iterative algorithm for large scale general-form Tikhonov regularization
- Binned Multilevel Monte Carlo for Bayesian Inverse Problems with Large Data
- An efficient iterative approach for large-scale separable nonlinear inverse problems
- HyBR
- An approximate empirical Bayesian method for large-scale linear-Gaussian inverse problems
- Efficient generalized Golub-Kahan based methods for dynamic inverse problems
- A Bayesian Numerical Homogenization Method for Elliptic Multiscale Inverse Problems
- Hybrid projection methods with recycling for inverse problems
- Efficient Krylov subspace methods for uncertainty quantification in large Bayesian linear inverse problems
- Hybrid Projection Methods with Recycling for Inverse Problems
- Generalized Modes in Bayesian Inverse Problems
- Hybrid projection methods for large-scale inverse problems with mixed Gaussian priors
- Numerical linear algebra in data assimilation
- Tomographic inverse problems: mathematical challenges and novel applications. Abstracts from the workshop held April 30 -- May 5, 2023
- Computational methods for large-scale inverse problems: a survey on hybrid projection methods
- Hybrid Projection Methods for Solution Decomposition in Large-scale Bayesian Inverse Problems
- A preconditioned Krylov subspace method for linear inverse problems with general-form Tikhonov regularization
- Efficient iterative methods for hyperparameter estimation in large-scale linear inverse problems
- Flexible modified LSMR for least squares problems
- Iterative reconstruction methods for cosmological X-ray tomography
- Low-dimensional subspace regularization through structured tensor priors
- Sparsity-promoting hierarchical Bayesian model for EIT with a blocky target
- Projected Newton method for large-scale Bayesian linear inverse problems
- Augmented flexible Krylov subspace methods with applications to Bayesian inverse problems
- Distributed Tikhonov regularization for ill-posed inverse problems from a Bayesian perspective
- Characterizing GSVD by singular value expansion of linear operators and its computation
- Subspace projection regularization for large-scale Bayesian linear inverse problems
- Iterative Bayesian inversion with Gaussian mixtures: finite sample implementation and large sample asymptotics
This page was built for publication: Generalized Hybrid Iterative Methods for Large-Scale Bayesian Inverse Problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5372617)
