Machine Learning Based Forward Solver: An Automatic Framework in gprMax

Utsav Akhaury; Iraklis Giannakis; Craig Warren; Antonios Giannopoulos

doi:10.1109/iwagpr50767.2021.9843172

Machine Learning Based Forward Solver: An Automatic Framework in gprMax

Utsav Akhaury, Iraklis Giannakis, Craig Warren, Antonios Giannopoulos

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Citations (Scopus)

46 Downloads (Pure)

Abstract

General full-wave electromagnetic solvers, such as those utilizing the finite-difference time-domain (FDTD) method, are computationally demanding for simulating practical GPR problems. We explore the performance of a near-real-time, forward modeling approach for GPR that is based on a machine learning (ML) architecture. To ease the process, we have developed a framework that is capable of generating these ML-based forward solvers automatically. The framework uses an innovative training method that combines a predictive dimensionality reduction technique and a large data set of modeled GPR responses from our FDTD simulation software, gprMax. The forward solver is parameterized for a specific GPR application, but the framework can be extended in a straightforward manner to different electromagnetic problems.

Original language	English
Title of host publication	2021 11th International Workshop on Advanced Ground Penetrating Radar (IWAGPR)
Place of Publication	Piscataway
Publisher	IEEE
Number of pages	6
ISBN (Electronic)	9781665422536
ISBN (Print)	9781665446624
DOIs	https://doi.org/10.1109/iwagpr50767.2021.9843172
Publication status	Published - 1 Dec 2021

Keywords

Full-Waveform Inversion (FWI)
Machine Learning (ML)
Principle Component Analysis (PCA)
Singular Value Decomposition (SVD)
Random Forest
XGBoost (Extreme Gradient Boosting)

Access to Document

10.1109/iwagpr50767.2021.9843172

IWAGPR_GSoCAccepted author manuscript, 930 KB

Cite this

@inproceedings{2f53cfdd686648f98578cdf371d9275a,

title = "Machine Learning Based Forward Solver: An Automatic Framework in gprMax",

abstract = "General full-wave electromagnetic solvers, such as those utilizing the finite-difference time-domain (FDTD) method, are computationally demanding for simulating practical GPR problems. We explore the performance of a near-real-time, forward modeling approach for GPR that is based on a machine learning (ML) architecture. To ease the process, we have developed a framework that is capable of generating these ML-based forward solvers automatically. The framework uses an innovative training method that combines a predictive dimensionality reduction technique and a large data set of modeled GPR responses from our FDTD simulation software, gprMax. The forward solver is parameterized for a specific GPR application, but the framework can be extended in a straightforward manner to different electromagnetic problems.",

keywords = "Full-Waveform Inversion (FWI), Machine Learning (ML), Principle Component Analysis (PCA), Singular Value Decomposition (SVD), Random Forest, XGBoost (Extreme Gradient Boosting)",

author = "Utsav Akhaury and Iraklis Giannakis and Craig Warren and Antonios Giannopoulos",

note = "Funding information: The project was funded via the Google Summer of Code (GSoC) 2021 programme. GSoC initiative is a global program focused on bringing student developers into open source software development. The source code for this project can be found at https://github.com/gprMax/gprMax/pull/294 .",

year = "2021",

month = dec,

day = "1",

doi = "10.1109/iwagpr50767.2021.9843172",

language = "English",

isbn = "9781665446624",

booktitle = "2021 11th International Workshop on Advanced Ground Penetrating Radar (IWAGPR)",

publisher = "IEEE",

address = "United States",

}

TY - GEN

T1 - Machine Learning Based Forward Solver

T2 - An Automatic Framework in gprMax

AU - Akhaury, Utsav

AU - Giannakis, Iraklis

AU - Warren, Craig

AU - Giannopoulos, Antonios

N1 - Funding information: The project was funded via the Google Summer of Code (GSoC) 2021 programme. GSoC initiative is a global program focused on bringing student developers into open source software development. The source code for this project can be found at https://github.com/gprMax/gprMax/pull/294 .

PY - 2021/12/1

Y1 - 2021/12/1

N2 - General full-wave electromagnetic solvers, such as those utilizing the finite-difference time-domain (FDTD) method, are computationally demanding for simulating practical GPR problems. We explore the performance of a near-real-time, forward modeling approach for GPR that is based on a machine learning (ML) architecture. To ease the process, we have developed a framework that is capable of generating these ML-based forward solvers automatically. The framework uses an innovative training method that combines a predictive dimensionality reduction technique and a large data set of modeled GPR responses from our FDTD simulation software, gprMax. The forward solver is parameterized for a specific GPR application, but the framework can be extended in a straightforward manner to different electromagnetic problems.

AB - General full-wave electromagnetic solvers, such as those utilizing the finite-difference time-domain (FDTD) method, are computationally demanding for simulating practical GPR problems. We explore the performance of a near-real-time, forward modeling approach for GPR that is based on a machine learning (ML) architecture. To ease the process, we have developed a framework that is capable of generating these ML-based forward solvers automatically. The framework uses an innovative training method that combines a predictive dimensionality reduction technique and a large data set of modeled GPR responses from our FDTD simulation software, gprMax. The forward solver is parameterized for a specific GPR application, but the framework can be extended in a straightforward manner to different electromagnetic problems.

KW - Full-Waveform Inversion (FWI)

KW - Machine Learning (ML)

KW - Principle Component Analysis (PCA)

KW - Singular Value Decomposition (SVD)

KW - Random Forest

KW - XGBoost (Extreme Gradient Boosting)

UR - https://www.scopus.com/pages/publications/85136302918

U2 - 10.1109/iwagpr50767.2021.9843172

DO - 10.1109/iwagpr50767.2021.9843172

M3 - Conference contribution

SN - 9781665446624

BT - 2021 11th International Workshop on Advanced Ground Penetrating Radar (IWAGPR)

PB - IEEE

CY - Piscataway

ER -

Machine Learning Based Forward Solver: An Automatic Framework in gprMax

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this