Molecular modeling is an important topic in drug discovery. Decades of research have led to the development of high quality scalable molecular force fields. In this paper, we show that neural networks can be used to train a universal approximator for energy potential functions. By incorporating a fully automated training process we have been able to train smooth, differentiable, and predictive potential functions on large-scale crystal structures. A variety of tests have also been performed to show the superiority and versatility of the machine-learned model.
@article{arxiv.2103.04162,
title = {Molecular modeling with machine-learned universal potential functions},
author = {Ke Liu and Zekun Ni and Zhenyu Zhou and Suocheng Tan and Xun Zou and Haoming Xing and Xiangyan Sun and Qi Han and Junqiu Wu and Jie Fan},
journal= {arXiv preprint arXiv:2103.04162},
year = {2021}
}