FURI | Spring 2022
Using Machine Learned Forces and Energies to Improve Quantum Mechanical Simulations
Density Functional Theory (DFT) calculations have been used for decades to predict the chemical properties of compounds and materials. While DFT calculations are highly chemically accurate, they are computationally costly to perform for systems containing many atoms. On the other hand, classical mechanics-based molecular dynamics (MD) simulations can be performed at a larger scale and can model thousands of atoms, albeit with less chemical accuracy. Using machine learning to predict atomic forces and map atomic positions to an energy density will improve the accuracy of molecular dynamics simulations and will allow for faster prediction of material properties.
Student researcher
Matthew Hayes
Chemical engineering
Hometown: Tempe, Arizona, United States
Graduation date: Spring 2023