A Machine Learning Acceleration for Nonlinear PDE Solvers - applied to multiphase porous media flow

This repository is the official implementation of:

Machine learning acceleration for nonlinear solvers applied to multiphase porous media flow.

and

Online learning to accelerate nonlinear PDE solvers: applied to multiphase porous media flow (Submitted)

Directories:

• datasets: Train and test datasets.
• feature_selection: Grid search for different sets of features.
• model_selection: Grid search for different machine learning models.
• numerical_experiments: All numerical experiments.
• xgboost_coupling: XGBoost coupling.

Requirements

To install requirements:

 $ conda env create -f environment.yml 
 $ conda activate py3ml
 $ python -m ipykernel install --user --name=python3 (opitional)

Dataset

To download the dataset follow the instructions in the directory dataset.