Computational Bayesian inversion of operator equations with distributed uncertain input parameters is based on an infinite-dimensional version of Bayes’ formula established in M. Dashti and A.M. Stuart [Handbook of Uncertainty Quantification, edited by R. Ghanem, D. Higdon and H. Owhadi. Springer (2015).] and its numerical realization in C. Schillings and Ch. Schwab [Inverse Problems 29 (2013) 065011; Inverse Problems 30 (2014) 065007.] Based on the sparsity of the posterior density shown in C. Schillings and Ch. Schwab [Inverse Problems 29 (2013) 065011; Inverse Problems 30 (2014) 065007.]; C. Schwab and A.M. Stuart [Inverse Problems 28 (2012) 045003.], dimension-adaptive Smolyak quadratures can afford higher convergence rates than MCMC in terms of the number of solutions of the forward (parametric operator) equation in C. Schillings and Ch. Schwab [Inverse Problems 29 (2013) 065011; Inverse Problems 30 (2014) 065007.]. The error bounds and convergence rates obtained in C. Schillings and Ch. Schwab [Inverse Problems 29 (2013) 065011; Inverse Problems 30 (2014) 065007.] are independent of the parameter dimension (in particular free from the curse of dimensionality) but depend on the (co)variance of the additive, Gaussian observation noise as for some constant . It is proved that the Bayesian estimates admit asymptotic expansions as . Sufficient (nondegeneracy) conditions for the existence of finite limits as are presented. For Gaussian priors, these limits are shown to be related to MAP estimators obtained from Tikhonov regularized least-squares functionals. Quasi-Newton (QN) methods with symmetric rank-1 updates are shown to identify the concentration points in a non-intrusive way, and to obtain second order information of the posterior density at these points. Based on the theory, two novel computational Bayesian estimation algorithms for Bayesian estimation at small observation noise covariance with performance independent of are proposed: first, dimension-adaptive Smolyak quadrature from C. Schillings and Ch. Schwab [Inverse Problems 29 (2013) 065011; Inverse Problems 30 (2014) 065007.] combined with a reparametrization of the parametric Bayesian posterior density near the MAP point (assumed unique) and, second, generalized Richardson extrapolation to the limit of vanishing observation noise variance. Numerical experiments are presented which confirm -independent convergence of the curvature-rescaled, adaptive Smolyak algorithm. Dimension truncation of the posterior density is justified by a general compactness result for the posterior’s Hessian at the MAP point.
Accepted:
DOI: 10.1051/m2an/2016005
Mots-clés : Bayesian inverse problems, parametric operator equations, smolyak quadrature, sparsity, non-Gaussian prior, quasi-Newton methods, SR1 update, posterior reparametrization, richardson extrapolation
@article{M2AN_2016__50_6_1825_0, author = {Schillings, Claudia and Schwab, Christoph}, title = {Scaling limits in computational {Bayesian} inversion}, journal = {ESAIM: Mathematical Modelling and Numerical Analysis }, pages = {1825--1856}, publisher = {EDP-Sciences}, volume = {50}, number = {6}, year = {2016}, doi = {10.1051/m2an/2016005}, zbl = {1358.65013}, mrnumber = {3580124}, language = {en}, url = {http://www.numdam.org/articles/10.1051/m2an/2016005/} }
TY - JOUR AU - Schillings, Claudia AU - Schwab, Christoph TI - Scaling limits in computational Bayesian inversion JO - ESAIM: Mathematical Modelling and Numerical Analysis PY - 2016 SP - 1825 EP - 1856 VL - 50 IS - 6 PB - EDP-Sciences UR - http://www.numdam.org/articles/10.1051/m2an/2016005/ DO - 10.1051/m2an/2016005 LA - en ID - M2AN_2016__50_6_1825_0 ER -
%0 Journal Article %A Schillings, Claudia %A Schwab, Christoph %T Scaling limits in computational Bayesian inversion %J ESAIM: Mathematical Modelling and Numerical Analysis %D 2016 %P 1825-1856 %V 50 %N 6 %I EDP-Sciences %U http://www.numdam.org/articles/10.1051/m2an/2016005/ %R 10.1051/m2an/2016005 %G en %F M2AN_2016__50_6_1825_0
Schillings, Claudia; Schwab, Christoph. Scaling limits in computational Bayesian inversion. ESAIM: Mathematical Modelling and Numerical Analysis , Volume 50 (2016) no. 6, pp. 1825-1856. doi : 10.1051/m2an/2016005. http://www.numdam.org/articles/10.1051/m2an/2016005/
S. Agapiou, O. Papaspiliopoulos, D. Sanz-Alonso and A.M. Stuart, Importance Sampling: Computational Complexity and Intrinsic Dimension. Preprint (2015). | arXiv | MR
Resolution of Singularities and Division of Distributions. Comm. Pure Appl. Math. 23 (1970) 145–150. | DOI | MR | Zbl
,Analysis of the Hessian for Inverse Scattering Problems. Part II: Inverse Medium Scattering of Acoustic Waves. Inverse Problems 28 (2012) 055002. | DOI | MR | Zbl
and ,T. Bui-Thanh, C. Burstedde, O. Ghattas, J. Martin, G. Stadler and L.C. Wilcox, Extreme-scale UQ for Bayesian Inverse Problems Governed by PDE. ACM/IEEE Supercomputing SC12, Gordon Bell Prize Finalist. Salt Lake City, Utah (2012).
Strong and Weak Error Estimates for the Solutions of Elliptic PDEs with random coefficients. SIAM J. Numer. Anal. 50 (2013) 216–246. | DOI | MR | Zbl
,Sparse-Grid, Reduced-Basis Bayesian Inversion: Nonaffine-Parametric Nonlinear Equations. J. Comput. Phys. 316 (2016) 470–503. | DOI | MR | Zbl
and ,Breaking the Curse of Dimensionality in Sparse Polynomial Approximation of Parametric PDEs. J. Math. Pures Appliquées 103 (2015) 400–428. | DOI | MR | Zbl
, and ,Analytic Regularity and Polynomial Approximation of Parametric and Stochastic Elliptic PDEs. Anal. Appl. 9 (2011) 1–37. | DOI | MR | Zbl
, and ,Global Convergence of a Class of Trust Region Algorithms for Optimization with Simple Bounds. SIAM J. Numer. Anal. 25 (1988) 433–460. | DOI | MR | Zbl
, and ,Convergence of Quasi-Newton Matrices Generated by the Symmetric Rank-One Update. Math. Program. 50 (1991) 177–195. | DOI | MR | Zbl
, and ,Likelihood-Informed Dimension Reduction for Nonlinear Inverse Problems. Inverse Problems 30 (2014) 114015. | DOI | MR | Zbl
, , , and ,Dimension-Independent Likelihood-Informed MCMC. J. Comput. Phys. 304 (2016) 109–137. | DOI | MR | Zbl
, and ,M. Dashti and A.M. Stuart, The Bayesian Approach to Inverse Problems. Handbook of Uncertainty Quantification, edited by R. Ghanem, D. Higdon and H. Owhadi. Springer (2015). | MR
MAP Estimators and their Consistency in Bayesian Nonparametric Inverse Problems. Inverse Problems 29 (2013) 095017. | DOI | MR | Zbl
, , and ,Higher Order Quasi Monte Carlo Integration for Holomorphic, Parametric Operator Equations. J. Uncertainty. Quantification 4 (2016) 48–79. | DOI | MR
, and ,M.V. Fedoryuk, Asimptotika: integraly i ryady. Spravochnaya Matematicheskaya Biblioteka. [Mathematical Reference Library] (in russian). Nauka Publ., Moscow, Russia (1987). | Zbl
R. Fletcher, An Overview of Unconstrained Optimization in Algorithms for Continuous Optimization, edited by E. Spedicato. Springer, The Netherlands (1994) 109–14. | MR | Zbl
The Convergence of Variable Metric Matrices in Unconstrained Optimization. Math. Program. 27 (1983) 123–143. | DOI | MR | Zbl
and ,An Introduction to the Adjoint Approach to Design. Flow, Turbulence and Combustion 65 (2000) 393–415. | DOI | Zbl
and ,Riemann Manifold Langevin and Hamiltonian Monte Carlo Methods. J. R. Stat. Soc.: Series B 73 (2011) 123–214. | DOI | MR | Zbl
and ,Analytic Regularity and Best -Term Approximation of High-Dimensional Parametric Initial Value Problems. Vietnam J. Math. 41/2 (2012) 181–215. | MR | Zbl
and ,Analytic Regularity and Polynomial Approximation of Stochastic, Parametric Elliptic Multiscale PDEs. Anal. Appl. 11 (2013) 1350001. | DOI | MR | Zbl
and ,J. Kaipio and E. Somersalo, Statistical and Computational Inverse Problems. Vol. 160 of Appl. Math. Sci. Springer (2005). | MR | Zbl
Bayesian Inverse Problems with Gaussian Priors. Ann. Stat. 39 (2011) 2626–2657. | DOI | MR | Zbl
, and ,Proposals which speed up Function-Space MCMC. J. Comput. Appl. Math. 262 (2014) 127–138. | DOI | MR | Zbl
,Newton MCMC Method for Large-Scale Statistical Inverse Problems with Application to Seismic Inversion. SIAM J. Sci. Comput. 34 (2012) A1460–A1487. | DOI | MR | Zbl
, , and and ,Bayesian Inference with Optimal Maps. J. Comput. Phys. 231 (2012) 7815–7850. | DOI | MR | Zbl
and ,J. Nocedal and S.J. Wright, Numerical Optimization. Springer, 2nd Edition (1999). | MR | Zbl
Sparse, adaptive Smolyak Quadrature for Bayesian inverse problems. Inverse Problems 29 (2013) 065011. | DOI | MR | Zbl
and ,Sparsity in Bayesian inversion of parametric operator equations. Inverse Problems 30 (2014) 065007. | DOI | MR | Zbl
and ,C. Schillings and Ch. Schwab, Scaling Limits in Computational Bayesian Inversion. Report 2014-26, Seminar for Applied Mathematics.ETH Zürich, http://www.sam.math.ethz.ch (2014). | Numdam | MR
C. Schillings, M. Sunnåker, J. Stelling and Ch. Schwab, Efficient Characterization of Parametric Uncertainty of Complex (Bio)chemical Networks. PLOS Comp. Biol. (2015), . | DOI
Sparse deterministic approximation of Bayesian inverse problems. Inverse Problems 28 (2012) 045003. | DOI | MR | Zbl
and ,On a generalization of the Richardson extrapolation process. Numer. Math. 57 (1990) 365–377. | DOI | MR | Zbl
,Inverse problems: a Bayesian approach. Acta Numerica 19 (2010) 451–559. | DOI | MR | Zbl
,Accurate Approximations for Posterior Moments and Marginal Densities. J. Am. Stat. Assoc. 81 (1986) 82–86. | DOI | MR | Zbl
and ,Cited by Sources: