Quantum graph neural network (quantum GNN) for molecular property prediction

This code is a simpler model and its implementation of "Fast and Accurate Molecular Property Prediction: Learning Atomic Interactions and Potentials with Neural Networks (The Journal of Physical Chemistry Letters, 2018)" in PyTorch.

The learning curves (x-axis is the number of epochs (i.e., iterations) and y-axis is the error (MAE) in the unit of eV on the test dataset) are as follows:

These results can be reproduce by the two commands (see "Usage").

Characteristics

• This code is easy-to-use. The requirement is only PyTorch. Preprocessing a dataset and learning a model can be done by only two commands (see "Usage").

Requirements

• PyTorch

Usage

We provide two major scripts:

• code/preprocess_data.py creates the input tensor data of molecules for processing with PyTorch from the original data (see dataset/original/data.txt).
• code/run_training.py trains a quantum-GNN using the above preprocessed data to predict a molecular property.

(i) Create the tensor data of molecules and their properties with the following command:

cd code
bash preprocess_data.sh

(ii) Using the preprocessed data, train a quantum-GNN with the following command:

bash run_training.sh

The training result and trained model are saved in the output directory (after training, see output/result and output/model).

(iii) You can change the model hyperparameters in run_training.sh. Try to learn various models!

How to cite

@article{tsubaki2018fast,
  title={Fast and Accurate Molecular Property Prediction: Learning Atomic Interactions and Potentials with Neural Networks},
  author={Tsubaki, Masashi and Mizoguchi, Teruyasu},
  journal={The journal of physical chemistry letters},
  volume={9},
  number={19},
  pages={5733--5741},
  year={2018},
  publisher={ACS Publications}
}