A Machine Learning Framework for Modeling Ensemble Properties of Atomically Disordered Materials

Description

This work proposes a general statistical framework involving graph neural networks and Monte-Carlo simulations to predict the thermodynamic properties and ensemble-averged functional properties of disordered materials.

Dependencies

All required packages are listed in requirements.txt. Some key dependencies include:

• python==3.11
• torch==2.3.1
• torch_geometric==2.5.3
• e3nn==0.5.5

You can install them with:

pip install -r requirements.txt

Usage

For graph neural network training and testing, navigate to the GNN/ directory, and run the script:

python GNN.py

For Monte-Carlo simulation, nativate to the MC/ directory, and run the script:

python Metropolis.py

Citation

If you find this work useful, please consider cite the following reference: