MCMC methods for functions: modifying old algorithms to make them faster
From MaRDI portal
Abstract: Many problems arising in applications result in the need to probe a probability distribution for functions. Examples include Bayesian nonparametric statistics and conditioned diffusion processes. Standard MCMC algorithms typically become arbitrarily slow under the mesh refinement dictated by nonparametric description of the unknown function. We describe an approach to modifying a whole range of MCMC methods, applicable whenever the target measure has density with respect to a Gaussian process or Gaussian random field reference measure, which ensures that their speed of convergence is robust under mesh refinement. Gaussian processes or random fields are fields whose marginal distributions, when evaluated at any finite set of points, are -valued Gaussians. The algorithmic approach that we describe is applicable not only when the desired probability measure has density with respect to a Gaussian process or Gaussian random field reference measure, but also to some useful non-Gaussian reference measures constructed through random truncation. In the applications of interest the data is often sparse and the prior specification is an essential part of the overall modelling strategy. These Gaussian-based reference measures are a very flexible modelling tool, finding wide-ranging application. Examples are shown in density estimation, data assimilation in fluid mechanics, subsurface geophysics and image registration. The key design principle is to formulate the MCMC method so that it is, in principle, applicable for functions; this may be achieved by use of proposals based on carefully chosen time-discretizations of stochastic dynamical systems which exactly preserve the Gaussian reference measure. Taking this approach leads to many new algorithms which can be implemented via minor modification of existing algorithms, yet which show enormous speed-up on a wide range of applied problems.
Recommendations
- MCMC methods for sampling function space
- Proposals which speed up function-space MCMC
- A function space HMC algorithm with second order Langevin diffusion limit
- Signal processing problems on function space: Bayesian formulation, stochastic PDEs and effective MCMC methods
- Dimension-Independent MCMC Sampling for Inverse Problems with Non-Gaussian Priors
Cites work
- A First Course in the Numerical Analysis of Differential Equations
- A note on Metropolis-Hastings kernels for general state spaces
- Analysis of SPDEs arising in path sampling. I: The Gaussian case
- Analysis of SPDEs arising in path sampling. II: The nonlinear case
- Bayesian aspects of some nonparametric problems
- Bayesian data assimilation in shape registration
- Bayesian inverse problems for functions and applications to fluid mechanics
- Bayesian learning for neural networks
- Bayesian Nonparametrics
- Besov priors for Bayesian inverse problems
- Conditional path sampling of SDEs and the Langevin MCMC method
- Diffusion limits of the random walk Metropolis algorithm in high dimensions
- Exponential convergence of Langevin distributions and their discrete approximations
- Gaussian Markov Random Fields
- Group actions, homeomorphisms, and matching: A general framework
- scientific article; zbMATH DE number 1350773 (Why is no real title available?)
- scientific article; zbMATH DE number 472921 (Why is no real title available?)
- scientific article; zbMATH DE number 1124118 (Why is no real title available?)
- scientific article; zbMATH DE number 1983459 (Why is no real title available?)
- scientific article; zbMATH DE number 1522715 (Why is no real title available?)
- scientific article; zbMATH DE number 1827006 (Why is no real title available?)
- scientific article; zbMATH DE number 3250158 (Why is no real title available?)
- Hybrid Monte Carlo on Hilbert spaces
- Interpolation of spatial data. Some theory for kriging
- Inverse problems: a Bayesian perspective
- MCMC METHODS FOR DIFFUSION BRIDGES
- MCMC using Hamiltonian dynamics
- Monte Carlo strategies in scientific computing
- On inference for partially observed nonlinear diffusion models using the Metropolis-Hastings algorithm
- Optimal scaling and diffusion limits for the Langevin algorithm in high dimensions
- Optimal scaling for various Metropolis-Hastings algorithms.
- Optimal Scaling of Discrete Approximations to Langevin Diffusions
- Optimal tuning of the hybrid Monte Carlo algorithm
- Reversible jump MCMC for nonparametric drift estimation for diffusion processes
- Sampling conditioned diffusions
- Signal processing problems on function space: Bayesian formulation, stochastic PDEs and effective MCMC methods
- Spectral gaps for a Metropolis-Hastings algorithm in infinite dimensions
- Stochastic Equations in Infinite Dimensions
- The Geometry of Random Fields
- The variational particle-mesh method for matching curves
- Variational data assimilation using targetted random walks
- Weak convergence and optimal scaling of random walk Metropolis algorithms
Cited in
(only showing first 100 items - show all)- A Bayesian framework for the validation of models for subsurface flows: synthetic experiments
- Evaluation of Gaussian approximations for data assimilation in reservoir models
- Geometric MCMC for infinite-dimensional inverse problems
- Optimal strategies for the control of autonomous vehicles in data assimilation
- Hierarchical Bayesian level set inversion
- A modified randomized maximum likelihood for improved Bayesian history matching
- Efficient strategy for the Markov chain Monte Carlo in high-dimension with heavy-tailed target probability distribution
- Importance sampling: intrinsic dimension and computational cost
- On a generalization of the preconditioned Crank-Nicolson metropolis algorithm
- A Bayesian level set method for an inverse medium scattering problem in acoustics
- Fast sampling of parameterised Gaussian random fields
- Bayesian inference of random fields represented with the Karhunen-Loève expansion
- A dynamic bi-orthogonal field equation approach to efficient Bayesian inversion
- Posterior contraction for empirical Bayesian approach to inverse problems under non-diagonal assumption
- Bayesian inversion of a diffusion model with application to biology
- A Bayesian level set method for the shape reconstruction of inverse scattering problems in elasticity
- Conditional particle filters with diffuse initial distributions
- Ensemble sampler for infinite-dimensional inverse problems
- Generalized parallel tempering on Bayesian inverse problems
- Reconstruction, optimization, and design of heterogeneous materials and media: basic principles, computational algorithms, and applications
- Statistical guarantees for Bayesian uncertainty quantification in nonlinear inverse problems with Gaussian process priors
- Mixing rates for Hamiltonian Monte Carlo algorithms in finite and infinite dimensions
- Unbiased estimation using a class of diffusion processes
- Sticky PDMP samplers for sparse and local inference problems
- Bayesian updating and marginal likelihood estimation by cross entropy based importance sampling
- Continuum limit and preconditioned Langevin sampling of the path integral molecular dynamics
- Calibrate, emulate, sample
- Sequential ensemble transform for Bayesian inverse problems
- Solving and learning nonlinear PDEs with Gaussian processes
- Bayesian multiscale deep generative model for the solution of high-dimensional inverse problems
- Variational Bayesian approximation of inverse problems using sparse precision matrices
- A numerical method for solving linear systems in the preconditioned Crank-Nicolson algorithm
- Bayesian inversion for electrical-impedance tomography in medical imaging using the nonlinear Poisson-Boltzmann equation
- Comparison of regularized ensemble Kalman filter and tempered ensemble transform particle filter for an elliptic inverse problem with uncertain boundary conditions
- Hyperparameter estimation in Bayesian MAP estimation: parameterizations and consistency
- Efficient Bayesian shape-restricted function estimation with constrained Gaussian process priors
- Optimal scaling of random-walk Metropolis algorithms on general target distributions
- Sampling of Bayesian posteriors with a non-Gaussian probabilistic learning on manifolds from a small dataset
- An optimization problem based on a Bayesian approach for the 2D Helmholtz equation
- Localization for MCMC: sampling high-dimensional posterior distributions with local structure
- Adaptive dimension reduction to accelerate infinite-dimensional geometric Markov chain Monte Carlo
- Continuous-discrete smoothing of diffusions
- Accelerating uncertainty quantification of groundwater flow modelling using a deep neural network proxy
- A probabilistic finite element method based on random meshes: a posteriori error estimators and Bayesian inverse problems
- Multimodal Bayesian registration of noisy functions using Hamiltonian Monte Carlo
- Bayesian approach to inverse time-harmonic acoustic obstacle scattering with phaseless data generated by point source waves
- Proposals which speed up function-space MCMC
- The tamed unadjusted Langevin algorithm
- Spatial localization for nonlinear dynamical stochastic models for excitable media
- Random walk Metropolis algorithm in high dimension with non-Gaussian target distributions
- The Bayesian update: variational formulations and gradient flows
- Nonlocal TV-Gaussian prior for Bayesian inverse problems with applications to limited CT reconstruction
- Non-stationary phase of the MALA algorithm
- Solving Bayesian inverse problems from the perspective of deep generative networks
- Rejoinder: ``Probabilistic integration: a role in statistical computation?
- The parameter Houlihan: a solution to high-throughput identifiability indeterminacy for brutally ill-posed problems
- A Bayesian conjugate gradient method (with discussion)
- Dimension-independent likelihood-informed MCMC
- Optimization based methods for partially observed chaotic systems
- Cauchy difference priors for edge-preserving Bayesian inversion
- Error bounds for Metropolis-Hastings algorithms applied to perturbations of Gaussian measures in high dimensions
- A stable manifold MCMC method for high dimensions
- Efficient estimation of hydraulic conductivity heterogeneity with non-redundant measurement information
- Sparse approximation of triangular transports. I: The finite-dimensional case
- Bayesian neural network priors for edge-preserving inversion
- A unified performance analysis of likelihood-informed subspace methods
- Accelerated dimension-independent adaptive metropolis
- Variational data assimilation using targetted random walks
- On the convergence of adaptive sequential Monte Carlo methods
- Sampling constrained probability distributions using spherical augmentation
- Information-geometric Markov chain Monte Carlo methods using diffusions
- Signal processing problems on function space: Bayesian formulation, stochastic PDEs and effective MCMC methods
- A Randomized Maximum A Posteriori Method for Posterior Sampling of High Dimensional Nonlinear Bayesian Inverse Problems
- Uncertainty quantification in graph-based classification of high dimensional data
- Multilevel sequential Monte Carlo with dimension-independent likelihood-informed proposals
- Simulation of elliptic and hypo-elliptic conditional diffusions
- MALA-within-Gibbs samplers for high-dimensional distributions with sparse conditional structure
- Multilevel sequential importance sampling for rare event estimation
- The Random Feature Model for Input-Output Maps between Banach Spaces
- FEM-based discretization-invariant MCMC methods for PDE-constrained Bayesian inverse problems
- The Bayesian formulation of EIT: analysis and algorithms
- Efficient Adaptive MCMC Through Precision Estimation
- Algorithms for Kullback-Leibler approximation of probability measures in infinite dimensions
- Estimation of systematic and spatially correlated components of random signals from repeated measurements: application to contrast enhanced computer tomography measurements
- Bayesian approach to a nonlinear inverse problem for a time-space fractional diffusion equation
- How deep are deep Gaussian processes?
- Survey of multifidelity methods in uncertainty propagation, inference, and optimization
- Bayesian inversion in resin transfer molding
- Posterior consistency for Gaussian process approximations of Bayesian posterior distributions
- Iterative importance sampling algorithms for parameter estimation
- Iterative updating of model error for Bayesian inversion
- Random forward models and log-likelihoods in Bayesian inverse problems
- A hybrid adaptive MCMC algorithm in function spaces
- An Adaptive Independence Sampler MCMC Algorithm for Bayesian Inferences of Functions
- Spectral gaps for a Metropolis-Hastings algorithm in infinite dimensions
- Noisy gradient flow from a random walk in Hilbert space
- Determining white noise forcing from Eulerian observations in the Navier-Stokes equation
- Markov chain Monte Carlo for exact inference for diffusions
- An application of sparse measure valued Bayesian inversion to acoustic sound source identification
- On the consistency of graph-based Bayesian semi-supervised learning and the scalability of sampling algorithms
This page was built for publication: MCMC methods for functions: modifying old algorithms to make them faster
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5965033)
