1. De novo exploration and self-guided learning of potential-energy surfaces
- Author
-
Noam Bernstein, Volker L. Deringer, Gábor Csányi, Deringer, Volker L. [0000-0001-6873-0278], Apollo - University of Cambridge Repository, and Deringer, VL [0000-0001-6873-0278]
- Subjects
FOS: Physical sciences ,Interatomic potential ,02 engineering and technology ,01 natural sciences ,Computational science ,Molecular dynamics ,639/766/119/1002 ,0103 physical sciences ,lcsh:TA401-492 ,General Materials Science ,010306 general physics ,Protocol (object-oriented programming) ,Physics ,Flexibility (engineering) ,lcsh:Computer software ,639/301/1034/1037 ,Condensed Matter - Materials Science ,Artificial neural network ,article ,Materials Science (cond-mat.mtrl-sci) ,Computational Physics (physics.comp-ph) ,021001 nanoscience & nanotechnology ,cond-mat.mtrl-sci ,Computer Science Applications ,Range (mathematics) ,lcsh:QA76.75-76.765 ,Kernel (image processing) ,physics.comp-ph ,Mechanics of Materials ,Modeling and Simulation ,Metric (mathematics) ,lcsh:Materials of engineering and construction. Mechanics of materials ,0210 nano-technology ,Physics - Computational Physics - Abstract
Interatomic potential models based on machine learning (ML) are rapidly developing as tools for material simulations. However, because of their flexibility, they require large fitting databases that are normally created with substantial manual selection and tuning of reference configurations. Here, we show that ML potentials can be built in a largely automated fashion, exploring and fitting potential-energy surfaces from the beginning (de novo) within one and the same protocol. The key enabling step is the use of a configuration-averaged kernel metric that allows one to select the few most relevant and diverse structures at each step. The resulting potentials are accurate and robust for the wide range of configurations that occur during structure searching, despite only requiring a relatively small number of single-point DFT calculations on small unit cells. We apply the method to materials with diverse chemical nature and coordination environments, marking an important step toward the more routine application of ML potentials in physics, chemistry, and materials science.
- Published
- 2020
- Full Text
- View/download PDF