Use of convolutional neural networks in problems of recovery of the interatomic interaction potential for the classical molecular dynamics method

Автор: Zharikov Dmitry N., Zavyalov Dmitry V., Sivashova Ekaterina S.

Журнал: Сетевое научное издание «Системный анализ в науке и образовании» @journal-sanse

Статья в выпуске: 2, 2022 года.

Бесплатный доступ

The paper considers an approach to restoring the forces of interatomic interaction using a convolutional neural network in a model system of a two-dimensional gas of pointless structureless particles. The preliminary results of network training are presented. Keywords: Neural Network, Molecu

Neural network, molecular dynamics, interatomic potential

Короткий адрес: https://sciup.org/14124582

IDR: 14124582

Статья научная