U.S. Department of EnergyOffice of Scientific and Technical Information
Transfer learning of neural surrogates on multifidelity groundwater simulations
Abstract
Multifidelity data used in the paper published in Advances in Water Resources 206 (2025) 105140, https://doi.org/10.1016/j.advwatres.2025.105140 The code used to process the data is openly available on GitHub at https://github.com/Model-Reduction-and-UQ-Group/Transfer_Learning_K_reconstruction Computationally inexpensive surrogates of process-based models, such as deep neural networks, enable ensemble-based computations used in risk assessment, data assimilation, etc. However, generation of large datasets required to train a neural network can be as expensive as the ensemble simulations themselves. We ameliorate this challenge by using data from multifidelity (MF) groundwater simulations and transfer learning (TL) to reduce data generation costs while maintaining model accuracy. As a computational example, we train a deep convolutional neural network (CNN) to reconstruct permeability fields from saturation maps derived from a multiphase flow model. Starting with very low- and low-fidelity data generated on increasingly coarse meshes, we pretrain the CNN, followed by output-layer training and fine-tuning using only a limited number of high-fidelity samples. We demonstrate the surrogate’s robustness when interpreting low-quality inputs—such as interpolated maps or data affected by noise—which has strong implications for the applicability in practical hydrogeological scenarios. This multilevel MF-TL strategy achieves a favorable trade-off between computational efficiency and predictive accuracy, significantly outperforming high-fidelity-only approaches under the same computationalmore »
- Publication Date:
- DOE Contract Number:
- SC0023163
- Research Org.:
- Stanford University
- Sponsoring Org.:
- USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
- OSTI Identifier:
- 3005558
- DOI:
- https://doi.org/10.5281/zenodo.17087811
Citation Formats
Chiofalo, Alessia, Ciriello, Valentina, and Tartakovsky, Daniel M. Transfer learning of neural surrogates on multifidelity groundwater simulations. United States: N. p., 2024.
Web. doi:10.5281/zenodo.17087811.
Chiofalo, Alessia, Ciriello, Valentina, & Tartakovsky, Daniel M. Transfer learning of neural surrogates on multifidelity groundwater simulations. United States. doi:https://doi.org/10.5281/zenodo.17087811
Chiofalo, Alessia, Ciriello, Valentina, and Tartakovsky, Daniel M. 2024.
"Transfer learning of neural surrogates on multifidelity groundwater simulations". United States. doi:https://doi.org/10.5281/zenodo.17087811. https://www.osti.gov/servlets/purl/3005558. Pub date:Tue Dec 31 23:00:00 EST 2024
@article{osti_3005558,
title = {Transfer learning of neural surrogates on multifidelity groundwater simulations},
author = {Chiofalo, Alessia and Ciriello, Valentina and Tartakovsky, Daniel M.},
abstractNote = {Multifidelity data used in the paper published in Advances in Water Resources 206 (2025) 105140, https://doi.org/10.1016/j.advwatres.2025.105140 The code used to process the data is openly available on GitHub at https://github.com/Model-Reduction-and-UQ-Group/Transfer_Learning_K_reconstruction Computationally inexpensive surrogates of process-based models, such as deep neural networks, enable ensemble-based computations used in risk assessment, data assimilation, etc. However, generation of large datasets required to train a neural network can be as expensive as the ensemble simulations themselves. We ameliorate this challenge by using data from multifidelity (MF) groundwater simulations and transfer learning (TL) to reduce data generation costs while maintaining model accuracy. As a computational example, we train a deep convolutional neural network (CNN) to reconstruct permeability fields from saturation maps derived from a multiphase flow model. Starting with very low- and low-fidelity data generated on increasingly coarse meshes, we pretrain the CNN, followed by output-layer training and fine-tuning using only a limited number of high-fidelity samples. We demonstrate the surrogate’s robustness when interpreting low-quality inputs—such as interpolated maps or data affected by noise—which has strong implications for the applicability in practical hydrogeological scenarios. This multilevel MF-TL strategy achieves a favorable trade-off between computational efficiency and predictive accuracy, significantly outperforming high-fidelity-only approaches under the same computational budget.},
doi = {10.5281/zenodo.17087811},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Tue Dec 31 23:00:00 EST 2024},
month = {Tue Dec 31 23:00:00 EST 2024}
}