Close Search Advanced
Journals
Resources
About Us
Open Access

An Adaptive Surrogate Modeling Based on Deep Neural Networks for Large-Scale Bayesian Inverse Problems

An Adaptive Surrogate Modeling Based on Deep Neural Networks for Large-Scale Bayesian Inverse Problems
Preview
Add to basket

Year:    2020

Author:    Liang Yan, Tao Zhou

Communications in Computational Physics, Vol. 28 (2020), Iss. 5 : pp. 2180–2205

Abstract

In Bayesian inverse problems, surrogate models are often constructed to speed up the computational procedure, as the parameter-to-data map can be very expensive to evaluate. However, due to the curse of dimensionality and the nonlinear concentration of the posterior, traditional surrogate approaches (such us the polynomial-based surrogates) are still not feasible for large scale problems. To this end, we present in this work an adaptive multi-fidelity surrogate modeling framework based on deep neural networks (DNNs), motivated by the facts that the DNNs can potentially handle functions with limited regularity and are powerful tools for high dimensional approximations. More precisely, we first construct offline a DNN-based surrogate according to the prior distribution, and then, this prior-based DNN-surrogate will be adaptively & locally refined online using only a few high-fidelity simulations. In particular, in the refine procedure, we construct a new shallow neural network that views the previous constructed surrogate as an input variable – yielding a composite multi-fidelity neural network approach. This makes the online computational procedure rather efficient. Numerical examples are presented to confirm that the proposed approach can obtain accurate posterior information with a limited number of forward simulations.

Submit Article

You do not have full access to this article.

Already a Subscriber? Sign in as an individual or via your institution

Journal Article Details

Publisher Name:    Global Science Press

Language:    English

DOI:    https://doi.org/10.4208/cicp.OA-2020-0186

Communications in Computational Physics, Vol. 28 (2020), Iss. 5 : pp. 2180–2205

Published online:    2020-01

AMS Subject Headings:    Global Science Press

Copyright:    COPYRIGHT: © Global Science Press

Pages:    26

Keywords:    Bayesian inverse problems deep neural networks multi-fidelity surrogate modeling Markov chain Monte Carlo.

Author Details

Liang Yan

Tao Zhou

  1. Auto-weighted Bayesian Physics-Informed Neural Networks and robust estimations for multitask inverse problems in pore-scale imaging of dissolution

    Perez, Sarah | Poncet, Philippe

    Computational Geosciences, Vol. (2024), Iss.

    https://doi.org/10.1007/s10596-024-10313-x [Citations: 0]

  2. Randomized approaches to accelerate MCMC algorithms for Bayesian inverse problems

    Saibaba, Arvind K. | Prasad, Pranjal | de Sturler, Eric | Miller, Eric | Kilmer, Misha E.

    Journal of Computational Physics, Vol. 440 (2021), Iss. P.110391

    https://doi.org/10.1016/j.jcp.2021.110391 [Citations: 2]

  3. Gradient-free Stein variational gradient descent with kernel approximation

    Yan, Liang | Zou, Xiling

    Applied Mathematics Letters, Vol. 121 (2021), Iss. P.107465

    https://doi.org/10.1016/j.aml.2021.107465 [Citations: 4]

  4. Multifidelity data fusion in convolutional encoder/decoder networks

    Partin, Lauren | Geraci, Gianluca | Rushdi, Ahmad A. | Eldred, Michael S. | Schiavazzi, Daniele E.

    Journal of Computational Physics, Vol. 472 (2023), Iss. P.111666

    https://doi.org/10.1016/j.jcp.2022.111666 [Citations: 5]

  5. Stein variational gradient descent with local approximations

    Yan, Liang | Zhou, Tao

    Computer Methods in Applied Mechanics and Engineering, Vol. 386 (2021), Iss. P.114087

    https://doi.org/10.1016/j.cma.2021.114087 [Citations: 8]

  6. Bayesian Inversion with Neural Operator (BINO) for modeling subdiffusion: Forward and inverse problems

    Yan, Xiong-Bin | Xu, Zhi-Qin John | Ma, Zheng

    Journal of Computational and Applied Mathematics, Vol. 454 (2025), Iss. P.116191

    https://doi.org/10.1016/j.cam.2024.116191 [Citations: 0]

  7. Adaptive neural network surrogate model for solving the nonlinear elastic inverse problem via Bayesian inference

    Huo, Fuchang | Zhang, Kai | Gao, Yu | Li, Jingzhi

    Journal of Inverse and Ill-posed Problems, Vol. 0 (2024), Iss. 0

    https://doi.org/10.1515/jiip-2022-0050 [Citations: 0]

  8. Residual-based error correction for neural operator accelerated infinite-dimensional Bayesian inverse problems

    Cao, Lianghao | O'Leary-Roseberry, Thomas | Jha, Prashant K. | Oden, J. Tinsley | Ghattas, Omar

    Journal of Computational Physics, Vol. 486 (2023), Iss. P.112104

    https://doi.org/10.1016/j.jcp.2023.112104 [Citations: 8]

  9. Bayesian model error method for the passive inverse scattering problem

    Yin, Yunwen | Yan, Liang

    Inverse Problems, Vol. 40 (2024), Iss. 6 P.065005

    https://doi.org/10.1088/1361-6420/ad3f40 [Citations: 0]

  10. Goal-oriented a-posteriori estimation of model error as an aid to parameter estimation

    Jha, Prashant K. | Oden, J. Tinsley

    Journal of Computational Physics, Vol. 470 (2022), Iss. P.111575

    https://doi.org/10.1016/j.jcp.2022.111575 [Citations: 3]

  11. Sequential Model Correction for Nonlinear Inverse Problems

    Arjas, Arttu | Sillanpää, Mikko J. | Hauptmann, Andreas S.

    SIAM Journal on Imaging Sciences, Vol. 16 (2023), Iss. 4 P.2015

    https://doi.org/10.1137/23M1549286 [Citations: 0]

  12. A Bayesian Scheme for Reconstructing Obstacles in Acoustic Waveguides

    Gao, Yu | Liu, Hongyu | Wang, Xianchao | Zhang, Kai

    Journal of Scientific Computing, Vol. 97 (2023), Iss. 3

    https://doi.org/10.1007/s10915-023-02368-2 [Citations: 2]

  13. Inverse Modeling for Subsurface Flow Based on Deep Learning Surrogates and Active Learning Strategies

    Wang, Nanzhe | Chang, Haibin | Zhang, Dongxiao

    Water Resources Research, Vol. 59 (2023), Iss. 7

    https://doi.org/10.1029/2022WR033644 [Citations: 3]

  14. Bayesian inversion using adaptive Polynomial Chaos Kriging within Subset Simulation

    Rossat, D. | Baroth, J. | Briffaut, M. | Dufour, F.

    Journal of Computational Physics, Vol. 455 (2022), Iss. P.110986

    https://doi.org/10.1016/j.jcp.2022.110986 [Citations: 7]

  15. Scalable conditional deep inverse Rosenblatt transports using tensor trains and gradient-based dimension reduction

    Cui, Tiangang | Dolgov, Sergey | Zahm, Olivier

    Journal of Computational Physics, Vol. 485 (2023), Iss. P.112103

    https://doi.org/10.1016/j.jcp.2023.112103 [Citations: 7]

  16. Efficient multifidelity likelihood-free Bayesian inference with adaptive computational resource allocation

    Prescott, Thomas P. | Warne, David J. | Baker, Ruth E.

    Journal of Computational Physics, Vol. 496 (2024), Iss. P.112577

    https://doi.org/10.1016/j.jcp.2023.112577 [Citations: 1]

  17. Surrogate and inverse modeling for two-phase flow in porous media via theory-guided convolutional neural network

    Wang, Nanzhe | Chang, Haibin | Zhang, Dongxiao

    Journal of Computational Physics, Vol. 466 (2022), Iss. P.111419

    https://doi.org/10.1016/j.jcp.2022.111419 [Citations: 30]

  18. A deep neural network approach for parameterized PDEs and Bayesian inverse problems

    Antil, Harbir | Elman, Howard C | Onwunta, Akwum | Verma, Deepanshu

    Machine Learning: Science and Technology, Vol. 4 (2023), Iss. 3 P.035015

    https://doi.org/10.1088/2632-2153/ace67c [Citations: 3]

  19. Deep Importance Sampling Using Tensor Trains with Application to a Priori and a Posteriori Rare Events

    Cui, Tiangang | Dolgov, Sergey | Scheichl, Robert

    SIAM Journal on Scientific Computing, Vol. 46 (2024), Iss. 1 P.C1

    https://doi.org/10.1137/23M1546981 [Citations: 0]

  20. Surrogate modeling for Bayesian inverse problems based on physics-informed neural networks

    Li, Yongchao | Wang, Yanyan | Yan, Liang

    Journal of Computational Physics, Vol. 475 (2023), Iss. P.111841

    https://doi.org/10.1016/j.jcp.2022.111841 [Citations: 13]

  21. Image inversion and uncertainty quantification for constitutive laws of pattern formation

    Zhao, Hongbo | Braatz, Richard D. | Bazant, Martin Z.

    Journal of Computational Physics, Vol. 436 (2021), Iss. P.110279

    https://doi.org/10.1016/j.jcp.2021.110279 [Citations: 14]

  22. Multi-fidelity Bayesian neural networks: Algorithms and applications

    Meng, Xuhui | Babaee, Hessam | Karniadakis, George Em

    Journal of Computational Physics, Vol. 438 (2021), Iss. P.110361

    https://doi.org/10.1016/j.jcp.2021.110361 [Citations: 88]

  23. A MCMC method based on surrogate model and Gaussian process parameterization for infinite Bayesian PDE inversion

    Hu, Zheng | Wang, Hongqiao | Zhou, Qingping

    Journal of Computational Physics, Vol. 507 (2024), Iss. P.112970

    https://doi.org/10.1016/j.jcp.2024.112970 [Citations: 2]

  24. Adaptive weighting of Bayesian physics informed neural networks for multitask and multiscale forward and inverse problems

    Perez, Sarah | Maddu, Suryanarayana | Sbalzarini, Ivo F. | Poncet, Philippe

    Journal of Computational Physics, Vol. 491 (2023), Iss. P.112342

    https://doi.org/10.1016/j.jcp.2023.112342 [Citations: 8]