Bayesian Physics-Informed Neural Networks for the Subsurface Tomography Based on the Eikonal Equation

Rongxi Gou; Yijie Zhang; Xueyu Zhu; Jinghuai Gao

doi:10.1109/TGRS.2023.3286438

Back

Bayesian Physics-Informed Neural Networks for the Subsurface Tomography Based on the Eikonal Equation

Journal article

Peer reviewed

Bayesian Physics-Informed Neural Networks for the Subsurface Tomography Based on the Eikonal Equation

Rongxi Gou, Yijie Zhang, Xueyu Zhu and Jinghuai Gao

IEEE transactions on geoscience and remote sensing, Vol.61, 4503012

01/01/2023

DOI: 10.1109/TGRS.2023.3286438

View Online

Abstract

The high cost of acquiring a sufficient amount of seismic data for training has limited the use of machine learning in seismic tomography. In addition, the inversion uncertainty due to the noisy data and data scarcity is less discussed in the conventional seismic tomography literature. To mitigate the uncertainty effects and quantify their impacts in the prediction, the so-called Bayesian physics-informed neural networks (BPINNs) based on the eikonal equation are adopted to infer the velocity field and reconstruct the travel-time field. In BPINNs, two inference algorithms, including Stein variational gradient descent (SVGD) and Gaussian variational inference (VI), are investigated for the inference task. The numerical results of several benchmark problems demonstrate that the velocity field can be estimated accurately and the travel time can be well approximated with reasonable uncertainty estimates by BPINNs. This suggests that the inferred velocity model provided by BPINNs may serve as a valid initial model for seismic inversion and migration.

Algorithms

Machine Learning

Physics

Bayesian analysis

Bayesian theory

Eikonal equation

Inference

Inversion

Neural networks

Probability theory

Seismic data

Seismic surveys

Seismic tomography

Tomography

Travel

Travel time

Uncertainty

Velocity

Velocity distribution

Details

Title: Subtitle: Bayesian Physics-Informed Neural Networks for the Subsurface Tomography Based on the Eikonal Equation
Creators: Rongxi Gou - Xi'an Jiaotong University
Yijie Zhang - Xi'an Jiaotong University
Xueyu Zhu - University of Iowa
Jinghuai Gao - Xi'an Jiaotong University
Resource Type: Journal article
Publication Details: IEEE transactions on geoscience and remote sensing, Vol.61, 4503012
Publisher: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
DOI: 10.1109/TGRS.2023.3286438
ISSN: 0196-2892
eISSN: 1558-0644
Grant note: DOI: 10.13039/501100001809, name: National Natural Science Foundation of China, award: 42174137; DOI: 10.13039/100000893, name: Simons Foundation, award: 504054; DOI: 10.13039/501100012166, name: National Key Research and Development Program of China, award: 2020YFA0713400
Language: English
Date published: 01/01/2023
Academic Unit: Mathematics
Record Identifier: 9984438958402771

Metrics

1 Record Views

See more details