Skip over navigation

 

Transferability of Empirical Interatomic Potentials and the Knowledgebase of Interatomic Models

Daniel Karls (University of Minnesota), Ellad Tadmor (University of Minnesota), Ryan Elliott (University of Minnesota)

From Atomistics to Reality: Spanning Scales in Simulations and Experiments Symposium A

Wed 9:00 - 10:30

CIT 165

The problem of selecting an appropriate empirical interatomic potential for an application is one that has not been rigorously addressed by the materials simulation community. However, the forthcoming Knowledgebase of Interatomic Models (KIM) (https://www.openKIM.org) promises to serve as an abundant source of predictions from potentials and the corresponding first principles and experimental data for various material properties. Making use of this novel data resource in a cumulative manner, we employ a Bayesian, non-parametric supervised learning algorithm to systematically define and predict the transferability of empirical potentials. Our method facilitates the verification and understanding of complex experimental observations by identifying the most accurate potential to use in their investigation.