© Yu-Wei Chang, Laura Natali, Oveis Jamialahmadi, Stefano Romeo, Joana B. Pereira & Giovanni Volpe http://www.softmatterlab.org
GapNet is an alternative deep-learning training approach that can use highly incomplete datasets. This is the code for the arXiv preprint 2107.00429 Neural Network Training with Highly Incomplete Datasets.
- Python 3.8.5
- Tensorflow 2.5.0
- pydot 1.2.3
- Pandas 1.3.1
- scikit-learn 0.24.2
To see GapNet working principle, we provide two well-documented tutorial notebooks that train the GapNet model on a simulated dataset:
- gapnet_tutorial.ipynb demonstrates how to train a GapNet model on a simulated dataset with highly incomplete features.
- omparison_gapnet_vs_other_models.ipynb compares the performance between GapNet, and other models.
Each code example is a Jupyter Notebook that also includes detailed comments to guide the user. All neccesary files to run the code examples are provided.