EPSCoR Research Fellows: NSF: Developing a Deep Learning-based Multifunctional Method for Simulating Quantum Effects

Understanding how electrons, the particles carrying negative electric charge, move through materials is crucial for developing next-generation technologies. At the microscopic level, where electrons exhibit behaviors vastly different from those in everyday macroscopic objects, we encounter what are known as quantum effects. These effects are foundational to advances in modern physics and have led to innovations, such as superconductors, solar cells, and light-emitting diodes. By studying and harnessing these quantum effects, we can better understand the behaviors and dynamics of electrons in materials, which is essential for developing advanced devices. Moreover, recent advances in machine learning techniques offer new opportunities for creating powerful and efficient computational tools. This NSF EPSCoR Research Fellows project aims to develop new computational methods to simulate the fundamental behaviors and dynamics of electrons at the quantum level by incorporating cutting-edge machine-learning techniques. These simulations, based on an interdisciplinary approach, can push the boundaries of scientific knowledge and have the potential to deliver broad societal benefits through technological innovation.

This EPSCoR Research Infrastructure Improvement: EPSCoR Research Fellows project aims to provide a fellowship to an Assistant Professor and training for a graduate student at the University of Alabama at Birmingham (UAB). Unveiling the underlying mechanisms of quantum effects requires accurate simulations, with density functional theory (DFT) calculations serving as a powerful tool. However, as the system size increases, the computational cost of DFT calculations rises dramatically, making simulations increasingly expensive. Machine learning techniques show promising potential in addressing this challenge. By collaborating with an expert in quantum materials and method development at the host institute, the University of Texas at Austin, this project aims to develop a machine-learning-based, multifunctional computational method capable of simulating quantum effects in large-scale systems. This method will then be applied to investigate scientifically important systems, such as twisted 2D systems and 2D heterostructures. The research will involve methods of DFT, Wannier functions, and the state-of-the-art deep learning techniques. This proposed research not only broadens the research scope of the Principal Investigator (PI) to transform the PI's career trajectory, but also initiates multiple collaborations both within and between institutes. Furthermore, the research will result in open-source packages that will benefit the theoretical condensed matter physics community.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Award Number: 2428751
Principal Investigator: Yubo Qi
Funds Obligated: $256,892
State: AL