Conference ID: 969-339-25072
PIN Code: 311528
Accurate molecular simulation requires computationally expensive quantum chemistry models that makes simulating complex material phenomena or large molecules intractable. In this talk I will describe a new trend of “data-driven” multi-scale and coarse-grained models intended to overcome this barrier. I will focus in particular on numerical analysis aspects: high-dimensional approximation under symmetry constraints. Most current models in the literature rely on the power spectrum as a descriptor map to drive either ANN or GP fits. However, it can be shown that these descriptors are incomplete. An alternative approach is to construct a complete symmetric basis for old-fashioned linear regression.
Christoph Ortner is Professor of Mathematics at the University of Warwick, though he will move to UBC (Vancouver) in August 2020. After his D.Phil. (Ph.D.) in Numerical Analysis in Oxford (2006) CO stayed on as an RCUK Academic Fellow in Solid Mechanics and Mathematics of Materials. In 2011 CO joined the Warwick Mathematics Institute. CO’s interdisciplinary work spans applied analysis, numerical analysis, scientific computing and atomistic (material) modelling. A substantial component of his research to date has been the development of the mathematical theory of multi-scale methods (atomistic-to-continuum, QM/MM) for materials and in particular material defects. This work was recognized by a Philip Leverhulme Prize (2012), ERC Starting Grant (2012), Whitehead Prize (2015) the Oberwolfach John Todd Award (2017) and an ERC Consolidater Grant (2020). His main research interests now are mathematical aspects of data-driven coarse-graining methods.