Geometric MCMC for infinite-dimensional inverse problems
Bayesian inverse problems often involve sampling posterior distributions on infinite-dimensional function spaces. Traditional Markov chain Monte Carlo (MCMC) algorithms are characterized by deteriorating mixing times upon meshrefinement, when the finite-dimensional approximations become more accurat...
| Main Authors: | , , , , |
|---|---|
| Format: | Journal article |
| Published: |
Elsevier
2016
|
| _version_ | 1797058722685517824 |
|---|---|
| author | Beskos, A Girolami, M Lan, S Farrell, P Stuart, A |
| author_facet | Beskos, A Girolami, M Lan, S Farrell, P Stuart, A |
| author_sort | Beskos, A |
| collection | OXFORD |
| description | Bayesian inverse problems often involve sampling posterior distributions on infinite-dimensional function spaces. Traditional Markov chain Monte Carlo (MCMC) algorithms are characterized by deteriorating mixing times upon meshrefinement, when the finite-dimensional approximations become more accurate. Such methods are typically forced to reduce step-sizes as the discretization gets finer, and thus are expensive as a function of dimension. Recently, a new class of MCMC methods with mesh-independent convergence times has emerged. However, few of them take into account the geometry of the posterior informed by the data. At the same time, recently developed geometric MCMC algorithms have been found to be powerful in exploring complicated distributions that deviate significantly from elliptic Gaussian laws, but are in general computationally intractable for models defined in infinite dimensions. In this work, we combine geometric methods on a finite-dimensional subspace with mesh-independent infinite-dimensional approaches. Our objective is to speed up MCMC mixing times, without significantly increasing the computational cost per step (for instance, in comparison with the vanilla preconditioned Crank-Nicolson (pCN) method). This is achieved by using ideas from geometric MCMC to probe the complex structure of an intrinsic finite-dimensional subspace where most data information concentrates, while retaining robust mixing times as the dimension grows by using pCN-like methods in the complementary subspace. The resulting algorithms are demonstrated in the context of three challenging inverse problems arising in subsurface flow, heat conduction and incompressible flow control. The algorithms exhibit up to two orders of magnitude improvement in sampling efficiency when compared with the pCN method. |
| first_indexed | 2024-03-06T19:54:19Z |
| format | Journal article |
| id | oxford-uuid:250acaab-338f-4f49-826b-58dd48b931ff |
| institution | University of Oxford |
| last_indexed | 2024-03-06T19:54:19Z |
| publishDate | 2016 |
| publisher | Elsevier |
| record_format | dspace |
| spelling | oxford-uuid:250acaab-338f-4f49-826b-58dd48b931ff2022-03-26T11:53:29ZGeometric MCMC for infinite-dimensional inverse problemsJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:250acaab-338f-4f49-826b-58dd48b931ffSymplectic Elements at OxfordElsevier2016Beskos, AGirolami, MLan, SFarrell, PStuart, ABayesian inverse problems often involve sampling posterior distributions on infinite-dimensional function spaces. Traditional Markov chain Monte Carlo (MCMC) algorithms are characterized by deteriorating mixing times upon meshrefinement, when the finite-dimensional approximations become more accurate. Such methods are typically forced to reduce step-sizes as the discretization gets finer, and thus are expensive as a function of dimension. Recently, a new class of MCMC methods with mesh-independent convergence times has emerged. However, few of them take into account the geometry of the posterior informed by the data. At the same time, recently developed geometric MCMC algorithms have been found to be powerful in exploring complicated distributions that deviate significantly from elliptic Gaussian laws, but are in general computationally intractable for models defined in infinite dimensions. In this work, we combine geometric methods on a finite-dimensional subspace with mesh-independent infinite-dimensional approaches. Our objective is to speed up MCMC mixing times, without significantly increasing the computational cost per step (for instance, in comparison with the vanilla preconditioned Crank-Nicolson (pCN) method). This is achieved by using ideas from geometric MCMC to probe the complex structure of an intrinsic finite-dimensional subspace where most data information concentrates, while retaining robust mixing times as the dimension grows by using pCN-like methods in the complementary subspace. The resulting algorithms are demonstrated in the context of three challenging inverse problems arising in subsurface flow, heat conduction and incompressible flow control. The algorithms exhibit up to two orders of magnitude improvement in sampling efficiency when compared with the pCN method. |
| spellingShingle | Beskos, A Girolami, M Lan, S Farrell, P Stuart, A Geometric MCMC for infinite-dimensional inverse problems |
| title | Geometric MCMC for infinite-dimensional inverse problems |
| title_full | Geometric MCMC for infinite-dimensional inverse problems |
| title_fullStr | Geometric MCMC for infinite-dimensional inverse problems |
| title_full_unstemmed | Geometric MCMC for infinite-dimensional inverse problems |
| title_short | Geometric MCMC for infinite-dimensional inverse problems |
| title_sort | geometric mcmc for infinite dimensional inverse problems |
| work_keys_str_mv | AT beskosa geometricmcmcforinfinitedimensionalinverseproblems AT girolamim geometricmcmcforinfinitedimensionalinverseproblems AT lans geometricmcmcforinfinitedimensionalinverseproblems AT farrellp geometricmcmcforinfinitedimensionalinverseproblems AT stuarta geometricmcmcforinfinitedimensionalinverseproblems |