Atomic cluster expansion for materials modeling
Yury Lysogorskiy, Ruhr-Universität Bochum, Germany
The prediction of materials properties is critical for the development of many technologically relevant energy materials. On the atomistic scale many properties can be determined with high accuracy by density functional theory (DFT) calculations. However, computational costs limits high-throughput screening of new materials and calculations of more complex materials properties, such as thermodynamics.
These limitations can be overcome by machine learning interatomic potentials that are orders of magnitude faster than DFT but retains its accuracy. In this talk I will demonstrate a new class of interatomic potentials – atomic cluster expansion (ACE), that combines both machine learning and basis expansion approaches and show several examples of its applications.