Solving inverse problems using conditional invertible neural networks.

Solving inverse problems using conditional invertible neural networks. JCP ArXiv

Govinda Anantha Padmanabha, Nicholas Zabaras

PyTorch Implementation of Solving inverse problems using conditional invertible neural networks.

Highlights

• Rather than developing a surrogate for a forward model, we are training directly an inverse surrogate mapping output information of a physical system to an unknown input distributed parameter.
• A generative model based on conditional invertible neural networks (cINN) is developed.
• The cINN is trained to serve as an inverse surrogate model of physical systems governed by PDEs.
• The inverse surrogate model is used for the solution of inverse problems with unknown spatially-dependent parameters.
• The developed method is applied for the estimation of a non-Gaussian permeability field in multiphase flows using limited pressure and saturation data.

Inverse surrogate model:

Mapping: observations → input space

Dependencies

PyTorch 1.0.0
Python 3
H5py
Matplotlib
Numpy

Citation

If you find this GitHub repository useful for your work, please consider to cite this work:

@article{padmanabha2021solving,
title={Solving inverse problems using conditional invertible neural networks},
journal={Journal of Computational Physics},
pages={110194},
year={2021},
publisher={Elsevier}
doi = {https://doi.org/10.1016/j.jcp.2021.110194 },
url = {https://www.sciencedirect.com/science/article/pii/S0021999121000899},
author = {Govinda Anantha Padmanabha and Nicholas Zabaras}
}