PINNup: Robust neural network wavefield solutions using frequency upscaling and neuron splitting
Name:
JGR Solid Earth - 2022 - Huang - PINNup Robust neural network wavefield solutions using frequency upscaling and neuron.pdf
Size:
14.32Mb
Format:
PDF
Description:
Accepted Manuscript
Type
ArticleAuthors
Huang, Xinquan
Alkhalifah, Tariq Ali

KAUST Department
Earth Science and Engineering ProgramPhysical Science and Engineering (PSE) Division
Date
2022-06-10Permanent link to this record
http://hdl.handle.net/10754/672118
Metadata
Show full item recordAbstract
Seismic wave-equation based methods, e.g. full waveform inversion, are currently used to illuminate the interior of Earth. Solving for the frequency-domain scattered wavefield via physics-informed neural network (PINN) has great potential in increasing the flexibility and reducing the computational cost of seismic modeling and inversion. However, when dealing with high-frequency wavefields using PINN, its accuracy and training cost limit its application. Thus, we propose a novel implementation of PINN using frequency upscaling and neuron splitting, which allows the neural network model to grow in size as we increase the frequency while leveraging the information from the pre-trained model for lower-frequency wavefields, resulting in fast convergence to highly accurate wavefield solutions. Numerical results show that, compared to the commonly used PINN with random initialization, the proposed PINN exhibits notable superiority in terms of convergence and accuracy and can achieve neuron based high-frequency wavefield solutions with a shallow model.Citation
Huang, X., & Alkhalifah, T. (2022). PINNup: Robust neural network wavefield solutions using frequency upscaling and neuron splitting. Journal of Geophysical Research: Solid Earth. Portico. https://doi.org/10.1029/2021jb023703Sponsors
We thank KAUST for its support and the SWAG group for the collaborative environment. This work utilized the resources of the Supercomputing Laboratory at King Abdullah University of Science and Technology (KAUST) in Thuwal, Saudi Arabia.Publisher
American Geophysical Union (AGU)arXiv
2109.14536Additional Links
https://onlinelibrary.wiley.com/doi/10.1029/2021JB023703ae974a485f413a2113503eed53cd6c53
10.1029/2021jb023703