Speaker
Description
Many recent efforts have focused on developing accurate interatomic potentials via data driven approaches in the pursuit of accelerating scientific studies of atomistic mechanisms as well as discovery of new materials. However, the accuracy of the interatomic potentials is limited by the data on which these models are trained, often obtained from DFT calculations. Thus, improving the quality and scale of the ab-initio data remains a key bottleneck in pushing the accuracy of interatomic potentials.
In this talk, recent progress in addressing the system size and accuracy limitations of ab-initio materials simulations will be presented. In particular, the development of computational methods and numerical algorithms for conducting fast and accurate large-scale DFT calculations using adaptive finite-element discretization will be presented, which form the basis for the recently released DFT-FE open-source code. The computational efficiency, scalability and performance of DFT-FE will be presented, which demonstrates a significant outperformance of widely used plane-wave DFT codes. Some recent application studies that highlight the capabilities of DFT-FE will be presented. In improving the accuracy of DFT calculations towards quantum accuracy, recent progress in accurately solving the inverse DFT problem will be presented, which has enabled the computation of exact exchange-correlation potentials for polyatomic systems. Ongoing efforts on using the exact exchange-correlation potentials to develop a data-driven approach for improving the exchange-correlation functional description in DFT will be discussed.
