Sample MD trajectory

MD trajectory used for training graph convolutional neural networks as coarse grained force fields as described in the following publications:

• E. Ricci, G. Giannakopoulos, V. Karkaletsis, D. N. Theodorou, N. Vergadou. 2022. "Developing Machine-Learned Potentials for Coarse-Grained Molecular Simulations: Challenges and Pitfalls". In Proceedings of 12th Conference on Artificial Intelligence (SETN). ACM, New York, NY, USA, 7 pages. https://doi.org/10.1145/3549737.3549793. Open access https://zenodo.org/record/7078577
• Gerakinis, D.-P., Ricci, E., Giannakopoulos, G., Karkaletsis, V., Theodorou, D. N., & Vergadou, N. (2024). Machine Learning-Based Coarse Grained Interaction Potentials for Molecular Systems. Zenodo. https://doi.org/10.5281/zenodo.10501037

The compressed archive contains input and output files for a NVT molecular dynamics simulation of a system containing 500 molecules of liquid benzene at 300 K, performed using LAMMPS.

The reference code used to train the model, with detailed usage instructions, is available at: https://github.com/ml-multimem/schnetpack-for-bulk-systems