Nanoscale Materials Modelling Using Neural Networks

This paper presents the development of a neural network approach in conjunction with molecular dynamics simulations. Molecular dynamics encompasses limitations with regard to computational times required for fine grain simulations. Neural networks can be used as an efficient tool for broadening the computational envelope in parametric investigations of materials using molecular simulations. Here, this concept is validated for a molecular system with an applied side shear, consisting of 560 molecules surrounding a cylindrical void.