Accelerating Atomic-Scale Calculations of Material Properties using Machine Learning

12Aug - by aiuniverse - 0 - In Data Mining


Abstract: One of the most promising applications of machine learning to materials research is the development of atomic-scale models that can be used to accelerate the calculation of material properties by orders of magnitude relative to density functional theory with little loss of accuracy. To this end, I will demonstrate how machine learning, in the form of genetic programming, can be used to develop accurate and transferable many-body interatomic potential models that are as fast as the embedded atom method, making them suitable to model materials on extreme time and length scales.

The key to our approach is to explore a hypothesis space of models based on fundamental physical principles and to select models from this hypothesis space based on their accuracy, speed, and simplicity. We demonstrate our approach by developing fast and accurate interatomic potential models for copper that generalize well to properties they were not trained on. Our approach requires relatively small sets of training data, making it possible to generate training data using highly accurate methods at a reasonable computational cost.

I will also demonstrate how machine learning can be used to accelerate the construction of cluster expansions and how this approach can be used in the rational design of new catalysts. In particular, I will introduce surface structure and catalytic activity maps of alloy surfaces generated using lattice-parameter-dependent cluster expansions trained using a Bayesian method. I will demonstrate this approach by creating maps of the Pt-rich region of the Pt-Ni phase diagram, providing atomic-scale insights into the catalytic properties of Pt-Ni alloys. These maps will be used to illustrate the roles of atomic order and intermetallic phases in catalysis design.

Bio: Timothy Mueller is an assistant professor in the Department of Materials Science and Engineering at the Whiting School of Engineering, Johns Hopkins University. His research focuses on computational materials science, nanomaterials, materials informatics, and materials for energy storage and conversion.

Facebook Comments