Researcher, ITMO University (Saint-Petersburg, Russia)
29 Aug 2018, 14:00-15:00
SkolTech, TPOC-4, Nobel str. 1, Red Building, 3rd floor, Room 351
Special Seminar Guest: Grigory Kabatyansky
In spite of being a mature field of research studying of magnetism and magnetic materials remains one of the most exciting area in condensed matter physics. Indeed, tremendous progress in technological development over the last decades is mainly held by achievements in spintronics and related fields, stemming from pioneering works on giant and tunnel magnetoresistance and spin-transfer torque. In this talk, we discuss a general framework which allows a direct and self-contained evaluation of spin torques within microscopic theory. We further apply our findings to explore magnetization dynamics of magnetic Skyrmions (a whirling particle-like configuration of magnetization) in both anti- and ferromagnets and magnetic solitons, and provide a theoretical analysis of prototypical logical device based on such structures. In the meantime, successful applications of neural network algorithms and machine learning to big data analysis in various areas provoke the use of this methodology in condensed matter physics. In this talk, our recent findings on unsupervised learning with self-organizing maps will be discussed in the context of phase identification for magnetic materials (primarily, a model two-dimensional ferromagnet and bcc iron). We argue further integration of machine learning approach towards a better understanding of materials informatics.