Abstract
We report a new deep learning message passing network that takes inspiration from Newton's equations of motion to learn interatomic potentials and forces. With the advantage of directional information from trainable force vectors, and physics-infused operators that are inspired by Newtonian physics, the entire model remains rotationally equivariant, and many-body interactions are inferred by more interpretable physical features. We test NewtonNet on the prediction of several reactive and non-reactive high quality ab initio data sets including single small molecules, a large set of chemically diverse molecules, and methane and hydrogen combustion reactions, achieving state-of-the-art test performance on energies and forces with far greater data and computational efficiency than other deep learning models.
| Original language | English |
|---|---|
| Pages (from-to) | 333-343 |
| Number of pages | 11 |
| Journal | Digital Discovery |
| Volume | 1 |
| Issue number | 3 |
| DOIs | |
| State | Published - Jun 1 2022 |
Funding
The authors thank the National Institutes of Health for support under Grant No 5U01GM121667 for support of the machine learning method. We also thank the National Science Foundation under grant CHE-1955643 for support of the chemical application of hydrogen combustion. FHZ thanks the Research Foundation-Flanders (FWO) Postdoctoral Fellowship for support as a visiting Berkeley scholar. M. Liu thanks the China Scholarship Council for a visiting scholar fellowship. C. J. S. acknowledges funding by the Ministry of Innovation, Science and Research of North Rhine-Westphalia (\u201CNRW R\u00FCckkehrerprogramm\u201D) and an Early Postdoc. Mobility fellowship from the Swiss National Science Foundation. This research used computational resources of the National Energy Research Scienti\uE103c Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.