A machine learning based hybrid Multi-Fidelity Multi-Level Monte Carlo method for uncertainty quantification

Nagoor Kani Jabarullah Khan, Ahmed H. Elsheikh

Research output: Contribution to journalArticle

28 Downloads (Pure)

Abstract

This paper focuses on reducing the computational cost of the Monte Carlo method for uncertainty propagation. Recently, Multi-Fidelity Monte Carlo (MFMC) method [46, 48] and Multi-Level Monte Carlo (MLMC) method [44, 29] were introduced to reduce the computational cost of Monte Carlo method by making use of low- fidelity models that are cheap to an evaluation in addition to the high-fidelity models. In this paper, we use machine learning techniques to combine the features of both the MFMC method and the MLMC method into a single framework called Multi-Fidelity- Multi-Level Monte Carlo (MFML-MC) method. In MFML-MC method, we use a hierarchy of proper orthogonal decomposition (POD) based approximations of high- fidelity outputs to formulate a MLMC framework. Next, we utilize Gradient Boosted Tree Regressor (GBTR) to evolve the dynamics of POD based reduced order model (ROM) [54] on every level of the MLMC framework. Finally, we incorporate MFMC method in order to exploit the POD ROM as a level specific low-fidelity model in the MFML-MC method. We compare the performance of MFML-MC method with the Monte Carlo method that uses either a high-fidelity model or a single low-fidelity model on two subsurface flow problems with random permeability field. Numerical results suggest that MFML-MC method provides an unbiased estimator with speedups by orders of magnitude in comparison to Monte Carlo method that uses high-fidelity model only.

Original languageEnglish
Article number105
JournalFrontiers in Environmental Science
Volume7
DOIs
Publication statusPublished - 27 Aug 2019

Fingerprint

decomposition
machine learning
method
subsurface flow
cost
permeability
evaluation
comparison

Keywords

  • Machine learning
  • Multi-Fidelity Monte Carlo method
  • Multi-level Monte Carlo method
  • POD
  • Uncertainty quantification

ASJC Scopus subject areas

  • Environmental Science(all)

Cite this

@article{ea65e74b2f134132a348867cb73b3c6a,
title = "A machine learning based hybrid Multi-Fidelity Multi-Level Monte Carlo method for uncertainty quantification",
abstract = "This paper focuses on reducing the computational cost of the Monte Carlo method for uncertainty propagation. Recently, Multi-Fidelity Monte Carlo (MFMC) method [46, 48] and Multi-Level Monte Carlo (MLMC) method [44, 29] were introduced to reduce the computational cost of Monte Carlo method by making use of low- fidelity models that are cheap to an evaluation in addition to the high-fidelity models. In this paper, we use machine learning techniques to combine the features of both the MFMC method and the MLMC method into a single framework called Multi-Fidelity- Multi-Level Monte Carlo (MFML-MC) method. In MFML-MC method, we use a hierarchy of proper orthogonal decomposition (POD) based approximations of high- fidelity outputs to formulate a MLMC framework. Next, we utilize Gradient Boosted Tree Regressor (GBTR) to evolve the dynamics of POD based reduced order model (ROM) [54] on every level of the MLMC framework. Finally, we incorporate MFMC method in order to exploit the POD ROM as a level specific low-fidelity model in the MFML-MC method. We compare the performance of MFML-MC method with the Monte Carlo method that uses either a high-fidelity model or a single low-fidelity model on two subsurface flow problems with random permeability field. Numerical results suggest that MFML-MC method provides an unbiased estimator with speedups by orders of magnitude in comparison to Monte Carlo method that uses high-fidelity model only.",
keywords = "Machine learning, Multi-Fidelity Monte Carlo method, Multi-level Monte Carlo method, POD, Uncertainty quantification",
author = "{Kani Jabarullah Khan}, Nagoor and Elsheikh, {Ahmed H.}",
year = "2019",
month = "8",
day = "27",
doi = "10.3389/fenvs.2019.00105",
language = "English",
volume = "7",
journal = "Frontiers in Environmental Science",
issn = "2296-665X",
publisher = "Frontiers Media S. A.",

}

A machine learning based hybrid Multi-Fidelity Multi-Level Monte Carlo method for uncertainty quantification. / Kani Jabarullah Khan, Nagoor; Elsheikh, Ahmed H.

In: Frontiers in Environmental Science, Vol. 7, 105, 27.08.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A machine learning based hybrid Multi-Fidelity Multi-Level Monte Carlo method for uncertainty quantification

AU - Kani Jabarullah Khan, Nagoor

AU - Elsheikh, Ahmed H.

PY - 2019/8/27

Y1 - 2019/8/27

N2 - This paper focuses on reducing the computational cost of the Monte Carlo method for uncertainty propagation. Recently, Multi-Fidelity Monte Carlo (MFMC) method [46, 48] and Multi-Level Monte Carlo (MLMC) method [44, 29] were introduced to reduce the computational cost of Monte Carlo method by making use of low- fidelity models that are cheap to an evaluation in addition to the high-fidelity models. In this paper, we use machine learning techniques to combine the features of both the MFMC method and the MLMC method into a single framework called Multi-Fidelity- Multi-Level Monte Carlo (MFML-MC) method. In MFML-MC method, we use a hierarchy of proper orthogonal decomposition (POD) based approximations of high- fidelity outputs to formulate a MLMC framework. Next, we utilize Gradient Boosted Tree Regressor (GBTR) to evolve the dynamics of POD based reduced order model (ROM) [54] on every level of the MLMC framework. Finally, we incorporate MFMC method in order to exploit the POD ROM as a level specific low-fidelity model in the MFML-MC method. We compare the performance of MFML-MC method with the Monte Carlo method that uses either a high-fidelity model or a single low-fidelity model on two subsurface flow problems with random permeability field. Numerical results suggest that MFML-MC method provides an unbiased estimator with speedups by orders of magnitude in comparison to Monte Carlo method that uses high-fidelity model only.

AB - This paper focuses on reducing the computational cost of the Monte Carlo method for uncertainty propagation. Recently, Multi-Fidelity Monte Carlo (MFMC) method [46, 48] and Multi-Level Monte Carlo (MLMC) method [44, 29] were introduced to reduce the computational cost of Monte Carlo method by making use of low- fidelity models that are cheap to an evaluation in addition to the high-fidelity models. In this paper, we use machine learning techniques to combine the features of both the MFMC method and the MLMC method into a single framework called Multi-Fidelity- Multi-Level Monte Carlo (MFML-MC) method. In MFML-MC method, we use a hierarchy of proper orthogonal decomposition (POD) based approximations of high- fidelity outputs to formulate a MLMC framework. Next, we utilize Gradient Boosted Tree Regressor (GBTR) to evolve the dynamics of POD based reduced order model (ROM) [54] on every level of the MLMC framework. Finally, we incorporate MFMC method in order to exploit the POD ROM as a level specific low-fidelity model in the MFML-MC method. We compare the performance of MFML-MC method with the Monte Carlo method that uses either a high-fidelity model or a single low-fidelity model on two subsurface flow problems with random permeability field. Numerical results suggest that MFML-MC method provides an unbiased estimator with speedups by orders of magnitude in comparison to Monte Carlo method that uses high-fidelity model only.

KW - Machine learning

KW - Multi-Fidelity Monte Carlo method

KW - Multi-level Monte Carlo method

KW - POD

KW - Uncertainty quantification

UR - http://www.scopus.com/inward/record.url?scp=85069038341&partnerID=8YFLogxK

U2 - 10.3389/fenvs.2019.00105

DO - 10.3389/fenvs.2019.00105

M3 - Article

AN - SCOPUS:85069038341

VL - 7

JO - Frontiers in Environmental Science

JF - Frontiers in Environmental Science

SN - 2296-665X

M1 - 105

ER -