polyid 0.1.2.3

Methods to train message passing neural network models on polymer structures.

PolyID^TM provides a framework for building, training, and predicting polymer properities using graph neural networks. The codes leverages nfp, for building tensorflow-based message-passing neural networ, and m2p, for building polymer structures. The notebooks have been provided that demonstrate how to: (1) build polymer structures from a polymer database and split into a training/validation and test set, (2) train a message passing neural network from using the trainining/validation set, and (3) evaluate the trained network on the test set. These three notebooks follow the methodology used in the forthcoming publication.

1. Building polymer structures: examples/1_generate_polymer_structures.ipynb
2. Training a message passing neural network: examples/2_generate_and_train_models.ipynb
3. Predicting and evaluating a trained network: examples/3_evaluate_model_performance_and_DoV.ipynb

Additional notebooks have been provided to provide more examples and capabilities of the PolyID code base.

4. Checking domain of validity: examples/example_determine_domain-of-validity.ipynb
5. Generating hierarchical fingerprints for performance comparison: examples/example_hierarchical_fingerprints.ipynb
6. Predicting with the trained model: examples/example_predict_with_trained_models.ipynb

For more details, see the manuscript PolyID: Artificial Intelligence for Discovering Performance-Advantaged and Sustainable Polymers, Macromolecules 2023.

Cite

If you use PolyID in your work, please cite

@article{wilson2023polyid,
  title={PolyID: Artificial Intelligence for Discovering Performance-Advantaged and Sustainable Polymers},
  author={Wilson, A Nolan and St John, Peter C and Marin, Daniela H and Hoyt, Caroline B and Rognerud, Erik G and Nimlos, Mark R and Cywar, Robin M and Rorrer, Nicholas A and Shebek, Kevin M and Broadbelt, Linda J and Beckham, Gregg T and Crowley, Michael F},
  journal={Macromolecules},
  volume={56},
  number={21},
  pages={8547--8557},
  year={2023},
  publisher={ACS Publications}
}