Atom-centered machine-learning force field package

Li，Lei1; Ciufo，Ryan A.2; Lee，Jiyoung2; Zhou，Chuan1; Lin，Bo1; Cho，Jaeyoung2; Katyal，Naman2; Henkelman，Graeme2

2023-11-01

10.1016/j.cpc.2023.108883

Computer Physics Communications   Impact Factor & Quartile

0010-4655

1879-2944

In recent years, machine learning algorithms have been widely used for constructing force fields with an accuracy of ab initio methods and the efficiency of classical force fields. Here, we developed a python-based atom-centered machine-learning force field (PyAMFF) package to provide a simple and efficient platform for fitting and using machine learning force fields by implementing an atom-centered neural-network algorithm with Behler-Parrinello symmetry functions as structural fingerprints. The following three features are included in PyAMFF: (1) integrated Fortran modules for fast fingerprint calculations and Python modules for user-friendly integration through scripts and facile extension of future algorithms; (2) a pure Fortran backend to interface with the software, including the long-timescale dynamic simulation package EON, enabling both molecular dynamic simulations and adaptive kinetic Monte Carlo simulations with machine-learning force fields; and (3) integration with the Atomic Simulation Environment package for active learning and ML-based algorithm development. Here, we demonstrate an efficient parallelization of PyAMFF in terms of CPU and memory usage and show that the Fortran-based PyAMFF calculator exhibits a linear scaling relationship with the number of symmetry functions and the system size. Program summary: Program title: python-based atom-centered machine-learning force field (PyAMFF) CPC Library link to program files: https://doi.org/10.17632/fsn6dkcvrv.1 Developer's repository link: https://gitlab.com/pyamff/pyamff Licensing provisions: Apache License, 2.0 Nature of problem: Determine an approximate (surrogate) model based upon atomic forces and energies from density functional theory (DFT). With a surrogate model that is less computationally expensive to evaluate than DFT, there can be a rapid exploration of the potential energy surface, accelerated optimization to minima and saddle points, and ultimately, accelerated design of active materials where the kinetics are key to the material function. Solution method: The atomic environments of training data are calculated in terms of Behler-Parrinello fingerprints. These fingerprints are passed to a neural network which is trained to reproduce the energy and force of the training data. A parallel implementation and Fortran backend allow for efficient training and calculation of the resulting surrogate model. Examples of long-time simulations of materials on the surrogate model surfaces are provided.

Adaptive kinetic Monte Carlo Atomic force field Density functional theory Machine learning

SCI

English

First ; Corresponding

Guangdong Science and Technology Department[2021B1212040001];Guangdong Science and Technology Department[2021B1212040001];National Key Research and Development Program of China[2022YFA1503102];National Key Research and Development Program of China[2022YFA1503102];National Natural Science Foundation of China[22179058];National Natural Science Foundation of China[22179058];National Science Foundation[CHE-2102317];National Science Foundation[CHE-2102317];Welch Foundation[F-1841];Welch Foundation[F-1841];

Computer Science ; Physics

Computer Science, Interdisciplinary Applications ; Physics, Mathematical

WOS:001078446500001

ELSEVIER

PHYSICS

2-s2.0-85168992479

Scopus

1.Department of Materials Science and Engineering,Guangdong Provincial Key Laboratory of Functional Oxide Materials and Devices,Southern University of Science and Technology,Shenzhen,Guangdong,518055,China
2.Department of Chemistry,the Oden Institute for Computational Engineering and Sciences,University of Texas at Austin,Austin,78712-0231,United States

Li，Lei,Ciufo，Ryan A.,Lee，Jiyoung,et al. Atom-centered machine-learning force field package[J]. Computer Physics Communications,2023,292.

Li，Lei.,Ciufo，Ryan A..,Lee，Jiyoung.,Zhou，Chuan.,Lin，Bo.,...&Henkelman，Graeme.(2023).Atom-centered machine-learning force field package.Computer Physics Communications,292.

Li，Lei,et al."Atom-centered machine-learning force field package".Computer Physics Communications 292(2023).