Optimization-Based MCMC Methods for Nonlinear Hierarchical Statistical Inverse Problems

In many hierarchical inverse problems, not only do we want to estimate high- or infinite-dimensional model parameters in the parameter-to-observable maps, but we also have to estimate hyperparameters that represent critical assumptions in the statistical and mathematical modeling processes. As a joint effect of high-dimensionality, nonlinear dependence, and non-concave structures in the joint posterior posterior distribution over model parameters and hyperparameters, solving inverse problems in the hierarchical Bayesian setting poses a significant computational challenge. In this work, we aim to develop scalable optimization-based Markov chain Monte Carlo (MCMC) methods for solving hierarchical Bayesian inverse problems with nonlinear parameter-to-observable maps and a broader class of hyperparameters. Our algorithmic development is based on the recently developed scalable randomize-then-optimize (RTO) method [4] for exploring the high- or infinite-dimensional model parameter space. By using RTO either as a proposal distribution in a Metropolis-within-Gibbs update or as a biasing distribution in the pseudo-marginal MCMC [2], we are able to design efficient sampling tools for hierarchical Bayesian inversion. In particular, the integration of RTO and the pseudo-marginal MCMC has sampling performance robust to model parameter dimensions. We also extend our methods to nonlinear inverse problems with Poisson-distributed measurements. Numerical examples in PDE-constrained inverse problems and positron emission tomography (PET) are used to demonstrate the performance of our methods.

[1]  A. Doucet,et al.  Efficient implementation of Markov chain Monte Carlo when using an unbiased likelihood estimator , 2012, 1210.1871.

[2]  Johnathan M. Bardsley,et al.  Hierarchical regularization for edge-preserving reconstruction of PET images , 2010 .

[3]  F. Lucka.  Fast Markov chain Monte Carlo sampling for sparse Bayesian inference in high-dimensional inverse problems using L1-type priors , 2012, 1206.0262.

[4]  Christophe Andrieu,et al.  Establishing some order amongst exact approximations of MCMCs , 2014, 1404.6909.

[5]  C. Fox,et al.  Markov chain Monte Carlo Using an Approximation , 2005 .

[6]  W. Marsden.  I and J , 2012 .

[7]  James Martin,et al.  A Computational Framework for Infinite-Dimensional Bayesian Inverse Problems, Part II: Stochastic Newton MCMC with Application to Ice Sheet Flow Inverse Problems , 2013, SIAM J. Sci. Comput..

[8]  Marco A. Iglesias,et al.  Hierarchical Bayesian level set inversion , 2016, Statistics and Computing.

[9]  Christoph Schwab,et al.  Karhunen-Loève approximation of random fields by generalized fast multipole methods , 2006, J. Comput. Phys..

[10]  Andrew M. Stuart,et al.  Inverse problems: A Bayesian perspective , 2010, Acta Numerica.

[11]  James Martin,et al.  A Computational Framework for Infinite-Dimensional Bayesian Inverse Problems Part I: The Linearized Case, with Application to Global Seismic Inversion , 2013, SIAM J. Sci. Comput..

[12]  Faming Liang,et al.  Statistical and Computational Inverse Problems , 2006, Technometrics.

[13]  Daniela Calvetti,et al.  Hypermodels in the Bayesian imaging framework , 2008 .

[14]  Tsuyoshi Murata,et al.  {m , 1934, ACML.

[15]  C. Andrieu,et al.  The pseudo-marginal approach for efficient Monte Carlo computations , 2009, 0903.5480.

[16]  Zheng Wang,et al.  Bayesian Inverse Problems with l1 Priors: A Randomize-Then-Optimize Approach , 2016, SIAM J. Sci. Comput..

[17]  Kody J. H. Law.  Proposals which speed up function-space MCMC , 2014, J. Comput. Appl. Math..

[18]  Yaming Yu,et al.  To Center or Not to Center: That Is Not the Question—An Ancillarity–Sufficiency Interweaving Strategy (ASIS) for Boosting MCMC Efficiency , 2011 .

[19]  Nicholas Zabaras,et al.  Hierarchical Bayesian models for inverse problems in heat conduction , 2005 .

[20]  Nathan Halko,et al.  Finding Structure with Randomness: Probabilistic Algorithms for Constructing Approximate Matrix Decompositions , 2009, SIAM Rev..

[21]  Albert Tarantola,et al.  Inverse problem theory - and methods for model parameter estimation , 2004 .

[22]  G. Roberts,et al.  MCMC Methods for Functions: ModifyingOld Algorithms to Make Them Faster , 2012, 1202.0709.

[23]  Tiangang Cui,et al.  Semivariogram methods for modeling Whittle–Matérn priors in Bayesian inverse problems , 2018, Inverse Problems.

[24]  Andrew M. Stuart,et al.  Robust MCMC Sampling with Non-Gaussian and Hierarchical Priors in High Dimensions , 2018 .

[25]  Omar Ghattas,et al.  A Randomized Maximum A Posteriori Method for Posterior Sampling of High Dimensional Nonlinear Bayesian Inverse Problems , 2016, SIAM J. Sci. Comput..

[26]  Robert H. Halstead,et al.  Matrix Computations , 2011, Encyclopedia of Parallel Computing.

[27]  Zhewei Yao,et al.  A TV-Gaussian prior for infinite-dimensional Bayesian inverse problems and its numerical implementations , 2015, 1510.05239.

[28]  A. Wood,et al.  Simulation of Stationary Gaussian Processes in [0, 1] d , 1994 .

[29]  Tiangang Cui,et al.  A posteriori stochastic correction of reduced models in delayed‐acceptance MCMC, with application to multiphase subsurface inverse problems , 2018 .

[30]  Tim Hesterberg,et al.  Monte Carlo Strategies in Scientific Computing , 2002, Technometrics.

[31]  H. Rue,et al.  An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach , 2011 .

[32]  Johnathan M. Bardsley,et al.  MCMC-Based Image Reconstruction with Uncertainty Quantification , 2012, SIAM J. Sci. Comput..

[33]  Tiangang Cui,et al.  Dimension-independent likelihood-informed MCMC , 2014, J. Comput. Phys..

[34]  Tiangang Cui,et al.  Scalable Optimization-Based Sampling on Function Space , 2019, SIAM J. Sci. Comput..

[35]  M. Girolami,et al.  Hyperpriors for Matérn fields with applications in Bayesian inversion , 2016, Inverse Problems & Imaging.

[36]  Qingping Zhou,et al.  Bayesian inference and uncertainty quantification for image reconstruction with Poisson data , 2019, 1903.02075.

[37]  H. Haario,et al.  An adaptive Metropolis algorithm , 2001 .

[38]  Colin Fox,et al.  Fast Sampling in a Linear-Gaussian Inverse Problem , 2015, SIAM/ASA J. Uncertain. Quantification.

[39]  Andrew M. Stuart,et al.  Hyperparameter Estimation in Bayesian MAP Estimation: Parameterizations and Consistency , 2019, The SMAI journal of computational mathematics.

[40]  Daniel Rudolf,et al.  On a Generalization of the Preconditioned Crank–Nicolson Metropolis Algorithm , 2015, Found. Comput. Math..

[41]  Habib N. Najm,et al.  Dimensionality reduction and polynomial chaos acceleration of Bayesian inference in inverse problems , 2008, J. Comput. Phys..

[42]  Jere Heikkinen.  Statistical Inversion Theory in X-ray Tomography , 2008 .

[43]  Frances Y. Kuo,et al.  Fast random field generation with H-matrices , 2017, Numerische Mathematik.

[44]  Jun S. Liu,et al.  Sequential Monte Carlo methods for dynamic systems , 1997 .

[45]  Arvind K. Saibaba,et al.  Efficient Marginalization-Based MCMC Methods for Hierarchical Bayesian Inverse Problems , 2019, SIAM/ASA J. Uncertain. Quantification.

[46]  Andrew M. Stuart,et al.  Analysis of the Gibbs Sampler for Hierarchical Inverse Problems , 2013, SIAM/ASA J. Uncertain. Quantification.

[47]  G. Roberts,et al.  MCMC methods for diffusion bridges , 2008 .

[48]  Heikki Haario,et al.  Randomize-Then-Optimize: A Method for Sampling from Posterior Distributions in Nonlinear Inverse Problems , 2014, SIAM J. Sci. Comput..

[49]  Andrew T. A. Wood,et al.  Algorithm AS 312: An Algorithm for Simulating Stationary Gaussian Random Fields , 1997 .

[50]  Tiangang Cui,et al.  Bayesian calibration of a large‐scale geothermal reservoir model by a new adaptive delayed acceptance Metropolis Hastings algorithm , 2011 .

[51]  Matthias Morzfeld,et al.  A random map implementation of implicit filters , 2011, J. Comput. Phys..

[52]  Harri Hakula,et al.  Conditionally Gaussian Hypermodels for Cerebral Source Localization , 2008, SIAM J. Imaging Sci..

[53]  James Martin,et al.  A Stochastic Newton MCMC Method for Large-Scale Statistical Inverse Problems with Application to Seismic Inversion , 2012, SIAM J. Sci. Comput..

[54]  Frances Y. Kuo,et al.  Analysis of Circulant Embedding Methods for Sampling Stationary Random Fields , 2017, SIAM J. Numer. Anal..

[55]  G. Roberts,et al.  On inference for partially observed nonlinear diffusion models using the Metropolis–Hastings algorithm , 2001 .

[56]  Hermann G. Matthies,et al.  Application of hierarchical matrices for computing the Karhunen–Loève expansion , 2009, Computing.

[57]  Leonhard Held,et al.  Gaussian Markov Random Fields: Theory and Applications , 2005 .

[58]  Dean S. Oliver,et al.  Metropolized Randomized Maximum Likelihood for Improved Sampling from Multimodal Distributions , 2015, SIAM/ASA J. Uncertain. Quantification.