An Information-Theoretic Approach to Multiscale Materials Modeling and Design Using High-throughput data
Nicholas Zabaras
Viola D. Hank Professor of Computational Science and Engineering
Institute for Computational Science and Informatics
University of Notre Dame, IN
ABSTRACT
We will introduce a number of fundamental themes in the analysis and design of materials in the presence of uncertainties using high-throughput data of structure and properties. We view uncertainty in a broad sense accounting for model inadequacy and model error (phenomenology typical of most materials models), lack of information due to limited training data-driven modeling, and loss of information when linking scales from electronic structure to continuum. We advocate a Bayesian information-theoretic perspective for modeling of these problems. We will present the fundamental problems that need to be addressed in an attempt to model materials under uncertainty. They include uncertainty modeling in density functional theory due to assumed exchange correlation functionals, predictive modeling of thermodynamic and thermomechanical properties, predictive coarse-graining in atomistic and molecular simulations, selection of the most informative materials simulations for particular quantities of interest in materials design, modeling of random polycrystalline materials, exploring property/structure/process relations under uncertainty and other. We will finally propose an inverse problem approach to materials design to allow design of materials with extremal properties. Such approaches are fundamentally different from combinatorial approaches that simply search an existing database of materials and properties with the hope of discovering a material with the desired properties.
DOI：10.12110/firstfmge.20171121.102
Nicholas Zabaras received his Diploma Degree in Mechanical Engineering at the National Technical University of Athens, Greece (1982), a M.S. in Materials at the University of Rochester, NY (1983) and PhD at Cornell University (1987) in Theoretical and Applied Mechanics. Upon receiving his Ph.D., he joined the faculty of Engineering at the University of Minnesota, Minneapolis, MN. Early research focused on the solution of inverse/design problems in the area of materials processing. In 1991, he returned to Cornell as a faculty of the Sibley School of Mechanical and Aerospace Engineering. At Cornell, he was also member of several other academic fields including Materials Science and Engineering, Applied Mathematics, Computational Science and Engineering. He was the founding director of the Materials Process Design and Control Laboratory (MPDC) that emphasized innovative materials modelling and design research with methodological approaches in mathematics, statistics and scientific computing. Particular areas with major contributions included inverse problems, uncertainty quantification and multiscale/multiphysics modelling. In the summer of 2014, he joined the University of Warwick as the Chair of Uncertainty Quantification to become the founding director of the Warwick Centre for Predictive Modelling (WCPM). WCPM is a university wide initiative across many colleges and departments that emphasizes integration of computational mathematics, computational statistics and scientific computing to address modelling and design of complex multiscale/multiphysics systems in the presence of uncertainties. In 2016, Prof. Zabaras took the title of Viola D. Hank Professor of computational science and engineering, at the University of Norte Dame, the and has worked there ever since. Prof. Zabaras has received several awards for his work. They include a 1991 Presidential Young Investigator Award for his work on Inverse Problems. In 2014, he was appointed as Hans Fisher Senior Fellow at the Institute of Advanced Study at the Technical University of Munich for his work on uncertainty quantification. At the same year he received the Royal Society's Wolfson Research Merit Award. He is Fellow and member of several professional societies. He currently serves as the Associate Editor of the Journal of Computational Physics and as the Editor in Chief of the International Journal for Uncertainty Quantification.