A seminar given by
Dr. Petr GRIGOREV
Research Fellow at Centre Interdisciplinaire de Nanoscience de Marseille (France)
Hybrid ab initio-machine learning simulations of extended defects
Abstract: In the context of plasma facing materials, interaction of radiation induced defects, plasma components and their clusters with material microstructure is important both in terms of material degradation as well as transport and retention of plasma components. Within this presentation I will describe how hybrid ab initio/machine learning methods can be used to study unfeasibly large systems with ab initio accuracy. I will show successful application of the method to study interaction of dislocations with hydrogen, helium and vacancy clusters in tungsten. I will also discuss future applications to fracture and grain boundary segregation and how this QM/ML data can be used to test and improve modern machine learning interatomic potentials.
Bio: Petr Grigorev is research Fellow working in the field of computational material science. He obtained his master degree in Physics from Peter the Great St. Petersburg Polytechnic University in 2012. The same year he enrolled in a Ph.D. program shared between Ghent University and Complutense University of Madrid within the framework of Erasmus Mundus FUSION-DC. He defended his Ph.D. in April 2017 and, shortly after that, joined Warwick Centre for Predictive Modelling as a Research Fellow. In December 2020 he started as a postdoc at the Départment Théorie et Simulation Numérique of Centre Interdisciplinaire de Nanoscience de Marseille (CINaM). Find more info here.
Photodissociation spectroscopy: a tool for probing biomolecules excited states properties