|About this Abstract
||2019 TMS Annual Meeting & Exhibition
||Computational Approaches for Big Data, Artificial Intelligence and Uncertainty Quantification in Computational Materials Science
||Modeling Complex Phenomena in 2D Materials Using First-principles Theory Based Machine Learning Force Fields
||Yang Yang, Hongxiang Zong, Hua Wang, Xiaodong Ding, Xiaofeng Qian
|On-Site Speaker (Planned)
Physical behaviors of materials are ultimately governed by complex interactions of electrons and ions, ranging from ultrafast process at atomistic scale to slow dynamics at mesoscale. Classical molecular dynamics is a powerful tool to understand the underlying mechanisms, however its application is limited by the availability of accurate force fields. In contrast, first-principles theory offers better accuracy, but the explicit description of electronic density and/or wavefunctions limits itself to short length and time scale. Here we present our recent development of first-principles theory based machine learning force fields for large-scale long-time simulations. We will show a few interesting examples of this method for 2D semimetals, insulators, and topological materials. The generated force fields are able to capture essential physics both qualitatively and quantitatively such as energetic, structural, and dynamic properties as well as complex phase transition, opening up unprecedented avenues for understanding and predicting complex physical behaviors of materials.
||Planned: Supplemental Proceedings volume