Energy-free machine learning force field for aluminum

Ivan Kruglov,Oleg Sergeev,Alexey Yanilkin,Artem R Oganov

doi:10.1038/s41598-017-08455-3

Abstract

We used the machine learning technique of Li et al. (PRL 114, 2015) for molecular dynamics simulations. Atomic configurations were described by feature matrix based on internal vectors, and linear regression was used as a learning technique. We implemented this approach in the LAMMPS code. The method was applied to crystalline and liquid aluminum and uranium at different temperatures and densities, and showed the highest accuracy among different published potentials. Phonon density of states, entropy and melting temperature of aluminum were calculated using this machine learning potential. The results are in excellent agreement with experimental data and results of full ab initio calculations.

Highlights

We used the machine learning technique of Li et al (PRL 114, 2015) for molecular dynamics simulations
Our tests show that the constructed Machine learning (ML) potentials can be used to reproduce the forces acting on atoms in the liquid state
The method is based on feature matrix description of atomic configurations and linear regression for the fitting

Summary

Method

The main parameters which should be optimized for ML potentials are the exact values of rcut and p pairs, and the number of such pairs and training set size. Since Al even with one optimally selected pair of parameters could be relatively well described (see Fig. 1a), all the main features of ML potential will be considered with reference to uranium α-phase (at zero pressure and 1000 K). Even though there are no explicitly calculated energies, a sufficiently accurate representation of the forces can enable the use of such potentials for modeling two-phase systems and for direct determination of the melting temperature. The obtained values are in reasonable agreement with the experimental melting temperature of 933 K and the equilibrium atomic volume for liquid of 18.9 Å3 Calculated pressure with correction is close to normal conditions

Conclusions

Findings

Additional Information