High-dimensional nonlinear Bayesian inference of poroelastic fields from pressure data

被引:2
|
作者
Karimi, Mina [1 ]
Massoudi, Mehrdad [2 ]
Dayal, Kaushik [1 ,3 ,4 ,5 ]
Pozzi, Matteo [1 ,5 ]
机构
[1] Carnegie Mellon Univ, Dept Civil & Environm Engn, Pittsburgh, PA 15213 USA
[2] Natl Energy Technol Lab, Pittsburgh, PA USA
[3] Carnegie Mellon Univ, Ctr Nonlinear Anal, Dept Math Sci, Pittsburgh, PA 15213 USA
[4] Carnegie Mellon Univ, Dept Mech Engn, Pittsburgh, PA 15213 USA
[5] Carnegie Mellon Univ, Wilton E Scott Inst Energy Innovat, Pittsburgh, PA 15213 USA
基金
美国国家科学基金会; 美国安德鲁·梅隆基金会;
关键词
Hamiltonian Monte Carlo; high-dimensional inference; Markov Chain Monte Carlo; poroelastic model; HAMILTONIAN MONTE-CARLO; STOCHASTIC NEWTON MCMC; INVERSE PROBLEMS; COMPUTATIONAL FRAMEWORK; MODEL-REDUCTION; INJECTION; OKLAHOMA; LANGEVIN;
D O I
10.1177/10812865221140840
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
We investigate solution methods for large-scale inverse problems governed by partial differential equations (PDEs) via Bayesian inference. The Bayesian framework provides a statistical setting to infer uncertain parameters from noisy measurements. To quantify posterior uncertainty, we adopt Markov Chain Monte Carlo (MCMC) approaches for generating samples. To increase the efficiency of these approaches in high-dimension, we make use of local information about gradient and Hessian of the target potential, also via Hamiltonian Monte Carlo (HMC). Our target application is inferring the field of soil permeability processing observations of pore pressure, using a nonlinear PDE poromechanics model for predicting pressure from permeability. We compare the performance of different sampling approaches in this and other settings. We also investigate the effect of dimensionality and non-gaussianity of distributions on the performance of different sampling methods.
引用
收藏
页码:2108 / 2131
页数:24
相关论文
共 50 条
  • [31] Bayesian clinical classification from high-dimensional data: Signatures versus variability
    Shalabi, Akram
    Inoue, Masato
    Watkins, Johnathan
    De Rinaldis, Emanuele
    Coolen, Anthony C. C.
    STATISTICAL METHODS IN MEDICAL RESEARCH, 2018, 27 (02) : 336 - 351
  • [32] Learning Discriminative Bayesian Networks from High-Dimensional Continuous Neuroimaging Data
    Zhou, Luping
    Wang, Lei
    Liu, Lingqiao
    Ogunbona, Philip
    Shen, Dinggang
    IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2016, 38 (11) : 2269 - 2283
  • [33] Change-point inference for high-dimensional heteroscedastic data
    Wu, Teng
    Volgushev, Stanislav
    Shao, Xiaofeng
    ELECTRONIC JOURNAL OF STATISTICS, 2023, 17 (02): : 3893 - 3941
  • [34] Model-Free Statistical Inference on High-Dimensional Data
    Guo, Xu
    Li, Runze
    Zhang, Zhe
    Zou, Changliang
    JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2024,
  • [35] Inference for mixed models of ANOVA type with high-dimensional data
    Chen, Fei
    Li, Zaixing
    Shi, Lei
    Zhu, Lixing
    JOURNAL OF MULTIVARIATE ANALYSIS, 2015, 133 : 382 - 401
  • [36] Online Variational Bayes Inference for High-Dimensional Correlated Data
    Kabisa, Sylvie
    Dunson, David B.
    Morris, Jeffrey S.
    JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS, 2016, 25 (02) : 426 - 444
  • [37] INFERENCE FOR CHANGE POINTS IN HIGH-DIMENSIONAL DATA VIA SELFNORMALIZATION
    Wang, Runmin
    Zhu, Changbo
    Volgushev, Stanislav
    Shao, Xiaofeng
    ANNALS OF STATISTICS, 2022, 50 (02): : 781 - 806
  • [38] On the challenges of learning with inference networks on sparse, high-dimensional data
    Krishnan, Rahul G.
    Liang, Dawen
    Hoffman, Matthew D.
    INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS, VOL 84, 2018, 84
  • [39] On inference in high-dimensional logistic regression models with separated data
    Lewis, R. M.
    Battey, H. S.
    BIOMETRIKA, 2024, 111 (03)
  • [40] Statistical inference via conditional Bayesian posteriors in high-dimensional linear regression
    Wu, Teng
    Narisetty, Naveen N.
    Yang, Yun
    ELECTRONIC JOURNAL OF STATISTICS, 2023, 17 (01): : 769 - 797