The Development of integrated machine-learning molecular dynamics system and its applications

Okumura, Masahiko   

Machine-learning molecular dynamics (MLMD) is an atomistic-scale simulation method that has achieved high accuracy and low computational cost, attracting attention in the nuclear energy research field. However, it has yet to be available to everyone. We are developing an integrated MLMD system that enables anyone to perform MLMD simulations. In this presentation, the presenter will give an overview of the system, its development progress, and how to use it in practice. Furthermore, the MLMD simulation of amorphous ice, a difficult material to handle in simulations because it is an amorphous state of polar molecules, will be shown as an example.