HetSys-WCPM Keynote Seminar: Building Useful Machine-Learned Interatomic Potentials
Brigham Young University
Monday 30th May 2022
1.00pm - 2.00pm, D2.02 School of Engineering, University of Warwick
Venue: OC0.01 Oculus Building, University of Warwick
Interatomic Potentials have long been used for atomistic modeling where the interesting questions are out of reach by first-principles approaches. Traditional empirical potentials are typically fitted to experimental data. They typically have poor general accuracy but are physically well-behaved. On the other hand, machine-learned interatomic potentials are far more expressive than physically motivated interatomic potentials like Lennard-Jones, Stillinger-Weber, Embedded Atom Potentials, etc., but they are also more likely to be completely wrong outside of the training domain, are more difficult to train reliably, and are computationally expensive. We have developed MLIPs for the Hf-Ni-Ti shape memory alloy. We share cautionary tales, best practices for generating training sets, and demonstrate how community tools make for "easy entry" to realistic thermodynamic modeling with these potentials.
About Professor Gus Hart:
I study materials physics. I want to help change the world by inventing algorithms for discovering the materials of tomorrow, today. I am a professor in the Department of Physics and Astronomy at Brigham Young University (BYU). I also serve as an Associate Dean in the College of Physical and Mathematical Sciences. Before coming to BYU, I was an assistant professor at Northern Arizona University (NAU). Prior to my academic appointments, I worked in the Solid State Theory Group with Alex Zunger at the National Renewable Energy Laboratory. I received a PhD from Univ. of California, Davis under Barry M. Klein.
My research focus is machine learning (ML) for materials discovery. This includes "high-throughput" computational materials science, developing ML algorithms and crystal structure representations, and generating new mathematics for modeling. I am a co-developer of the UNCLE code for cluster expansion modeling and a member of the aflow.org consortium. I benefit from collaborations with fantastic mathematicians, computer scientists, chemists, and fellow physicists at BYU and all over the world.