Seminar: Topological Material Database and Machine Learning On It

ABSTRACT

The establishment of topological band theory has made the topological classification of electronic states of solid materials increasingly rich and complete. An effective model based on the energy band inversion reveals its microscopic mechanism. Symmetry analysis of Bloch wave function provides direct and effective tools for identifying the types of topological states, known as the symmetry indicator theory and topological quantum chemistry theory. High-throughput electronic structure calculations can screen all known crystal materials and establish a topological materials database, which greatly promotes the discovery of topological materials. But how to turn the physical  mechanisms and empirical parameters into intuitive chemical images and material experiences is crucial to exploring and designing new topological materials. Machine learning based on data analysis can discover and establish correlations between different physical quantities, and deepen our understanding of the structure, component and property of materials and their mutual influence. By analyzing tens of thousands of data in the topological materials database through machine learning, quite accurate topological classification of materials can be achieved. The most important features of crystal structural and component valence electrons are found to affect topological properties. These features are closely related to the crucial parameters in the effective model Hamiltonian, which provides guidance for finding and establishing the causal relationship between material chemistry experience and physical mechanism.

BIOGRAPHY

Professor of Physics at the Institute of Physics, Chinese Academy of Sciences, and director of the Condensed Matter Science Data Center, Chinese Academy of Sciences. He received his undergraduate degree from the Department of Physics of Nanjing University in 2000 and his PhD in condensed matter physics from Nanjing University in 2005. He works in the direction of computational condensed matter physics. In addition to the development of computational methods and programs, he focuses on the computational research of novel quantum phenomena of solids, including topological materials, diluted magnetic semiconductor and transition-metal compounds. His papers on theoretical prediction and experimental discovery of Weyl semimetals were selected into the collection to celebrate 125th anniversary of Physical Review journals by American Physical Society.