Machine Learning a General-Purpose Interatomic Potential for Silicon

A machine-learning based approach to computing the energy and forces of silicon atoms circumvents the need for complex electronic structure calculations, offering an efficient method for predicting material properties that can be extended to other materials.