Dr. Kyeongjae Cho received $995,707 from the Defense Advanced Research Projects Agency (DARPA) for his research on Machine Learning Methods for Predictive Atomic and Electronic Structures of Nanostructure Interfaces for THz Device Simulations with DFT Accuracy. This project addresses interatomic potentials and tight binding (TB) models that can be used to simulate thousands and tens of thousands of atoms with reasonable computing resources based on the density functional theory (DFT) calculations. Without relying on ad hoc fitting methods, Dr. Kyeongjae Cho’s study contributes to further research on developing interatomic potentials and TB models with DFT accuracy, which would allow the calculation of systems composed of thousands of atoms to be highly efficient and accurate.

In light of the successful development of machine learning interatomic potentials and TB models proposal, his research will also benefit the UTD NanoSim team’s representation of quantum jump in materials simulations based on the data science and artificial intelligence (AI) approach of machine learning techniques.