Multilevel Sequential Monte Carlo with Dimension-Independent Likelihood-Informed Proposals

被引:24
|
作者
Beskos, Alexandros [1 ]
Jasra, Ajay [2 ]
Law, Kody [3 ]
Marzouk, Youssef [4 ]
Zhou, Yan [2 ]
机构
[1] UCL, Dept Stat Sci, London, England
[2] NUS, Dept Stat & Appl Probabil, Singapore 117546, Singapore
[3] Univ Manchester, Sch Math, Manchester, Lancs, England
[4] MIT, Dept Aeronaut & Astronaut, Cambridge, MA 02139 USA
来源
SIAM-ASA JOURNAL ON UNCERTAINTY QUANTIFICATION | 2018年 / 6卷 / 02期
关键词
multilevel Monte Carlo; sequential Monte Carlo; Bayesian inverse problem; uncertainty quantification;
D O I
10.1137/17M1120993
中图分类号
O1 [数学];
学科分类号
0701 ; 070101 ;
摘要
In this article we develop a new sequential Monte Carlo method for multilevel Monte Carlo estimation. In particular, the method can be used to estimate expectations with respect to a target probability distribution over an infinite-dimensional and noncompact space-as produced, for example, by a Bayesian inverse problem with a Gaussian random field prior. Under suitable assumptions the MLSMC method has the optimal O(epsilon(-2)) bound on the cost to obtain a mean-square error of O(epsilon(2)). The algorithm is accelerated by dimension-independent likelihood-informed proposals [T. Cui, K. J. Law, and Y. M. Marzouk, (2016), J. Compd. Phys., 304, pp. 109-137] designed for Gaussian priors, leveraging a novel variation which uses empirical covariance information in lieu of Hessian information, hence eliminating the requirement for gradient evaluations. The efficiency of the algorithm is illustrated on two examples: (i) inversion of noisy pressure measurements in a PDE model of Darcy flow to recover the posterior distribution of the permeability field and (ii) inversion of noisy measurements of the solution of an SDE to recover the posterior path measure.
引用
收藏
页码:762 / 786
页数:25
相关论文
共 32 条
  • [21] SEQUENTIAL MONTE CARLO SAMPLERS FOR MARGINAL LIKELIHOOD COMPUTATION IN MULTIPLICATIVE EXPONENTIAL NOISE MODELS
    Dikmen, Onur
    Cemgil, A. Taylan
    2014 IEEE WORKSHOP ON STATISTICAL SIGNAL PROCESSING (SSP), 2014, : 276 - 279
  • [22] Estimating DSGE Models using Multilevel Sequential Monte Carlo in Approximate Bayesian Computation
    Alaminos, David
    Ramirez, Ana
    Fernandez-Gamez, Manuel A.
    Becerra-Vicario, Rafael
    JOURNAL OF SCIENTIFIC & INDUSTRIAL RESEARCH, 2020, 79 (01): : 21 - 25
  • [23] A multilevel homotopy MCMC sequential Monte Carlo filter for multi-target tracking
    Kang, Kai
    Maroulas, Vasileios
    Schizas, Ioannis D.
    Blasch, Erik
    2016 19TH INTERNATIONAL CONFERENCE ON INFORMATION FUSION (FUSION), 2016, : 2015 - 2021
  • [24] Multilevel sequential Monte Carlo: Mean square error bounds under verifiable conditions
    Del Moral, Pierre
    Jasra, Ajay
    Law, Kody J. H.
    STOCHASTIC ANALYSIS AND APPLICATIONS, 2017, 35 (03) : 478 - 498
  • [25] Sequential Monte Carlo methods for mixtures with normalized random measures with independent increments priors
    Griffin, J. E.
    STATISTICS AND COMPUTING, 2017, 27 (01) : 131 - 145
  • [26] Sequential Monte Carlo methods for mixtures with normalized random measures with independent increments priors
    J. E. Griffin
    Statistics and Computing, 2017, 27 : 131 - 145
  • [27] Multilevel dimension reduction Monte-Carlo simulation for high-dimensional stochastic models in finance
    Dang, Duy-Minh
    Xu, Qifan
    Wu, Shangzhe
    INTERNATIONAL CONFERENCE ON COMPUTATIONAL SCIENCE, ICCS 2015 COMPUTATIONAL SCIENCE AT THE GATES OF NATURE, 2015, 51 : 1583 - 1592
  • [28] Likelihood inference for Markov switching GARCH(1,1) models using sequential Monte Carlo
    Wee, Damien C. H.
    Chen, Feng
    Dunsmuir, William T. M.
    ECONOMETRICS AND STATISTICS, 2022, 21 : 50 - 68
  • [29] Quasi-sequential monte carlo visual tracking based on multilevel dynamic layer representations in confidence region
    Song T.
    Li O.
    Liu G.-Y.
    Tien Tzu Hsueh Pao/Acta Electronica Sinica, 2016, 44 (06): : 1355 - 1361
  • [30] Analysis of a Class of Multilevel Markov Chain Monte Carlo Algorithms Based on Independent Metropolis--Hastings\ast
    Madrigal-Cianci, Juan P.
    Nobile, Fabio
    Tempone, Raul
    SIAM-ASA JOURNAL ON UNCERTAINTY QUANTIFICATION, 2023, 11 (01): : 91 - 138