Using Fourier Neural Operator for Seismic Forward and Inverse Problems
Description:
We are presenting our work on using the Fourier Neural Operator to solve the wave equation to simulate the 3D ground motion simulations from earthquakes and use it in both forward and inverse problems. We use the U-Shaped Neural Operator (U-NO) to solve the Helmholtz equation to forward simulate the 3D seismic ground motions and use the trained model to invert the seismic source. The inversion results from synthetic data show promising results; we are currently working on the real-data inversion. We set up the physical domain in San Francisco Bay Area and use the past earthquakes for the inversion. We will report our preliminary inversion results in this presentation.
Session: New Frontiers in Seismic Observations and Modeling with Innovative Methods and Emerging Data on Earth and Other Planets [Poster]
Type: Poster
Date: 4/17/2026
Presentation Time: 08:00 AM (local time)
Presenting Author: Qingkai Kong
Student Presenter: No
Invited Presentation:
Poster Number: 114
Authors
Qingkai Kong Presenting Author Corresponding Author kongqk@berkeley.edu Lawrence Livermore National Laboratory |
Jiun-Ting Lin kong11@llnl.gov Lawrence Livermore National Laboratory |
Claire Doody kong11@llnl.gov Lawrence Livermore National Laboratory |
Caifeng Zou czou@caltech.edu California Institute of Technology |
Youngsoo Choi kong11@llnl.gov Lawrence Livermore National Laboratory |
Zachary Ross zross@caltech.edu California Institute of Technology |
Kamyar Azizzadenesheli kaazizzad@gmail.com Nvidia |
Arthur Rodgers rodgers7@llnl.gov Lawrence Livermore National Laboratory |
Robert Clayton rclayton@gmail.com California Institute of Technology |
Using Fourier Neural Operator for Seismic Forward and Inverse Problems
Category
New Frontiers in Seismic Observations and Modeling with Innovative Methods and Emerging Data on Earth and Other Planets