Machine Learning in Computational Materials Design

Kruglov, Ivan Alexandrovich
PhD in Condensed Matter Physics, Head of Computational Materials Design Laboratory, MIPT University,
Senior Research Fellow, Dukhov Automatics Research Institute (VNIIA),
Senior Research Fellow, XPANCEO, Moscow, Russia – Dubai, United Arab Emirates
ivan.kruglov@phystech.edu

Abstract
One of the most promising tasks of modern materials science is the development and improvement of computer modeling methods. The use of computational methods and machine learning methods can significantly speed up and simplify the tasks of searching for new materials with specified properties, as well as studying the properties of already known materials. A special place in this area of materials science is occupied by computer design of materials - a method of searching for new materials by chemical composition using modern methods of atomistic and quantum chemical modeling. With the development of computer-aided design of materials, it has become possible to determine the crystal structure of a material, which is the first and necessary step towards predicting its properties. This allows you to optimize the structure of the material and its chemical composition so that it is possible to control its properties.
In this talk, I will show the most recent results how machine learning can speed up the existing approaches in computational materials design, find new materials, & open new horizons in this field.

Keywords: computational materials science, machine learning