报告人：刘科 博士后 (德国慕尼黑大学)

时 间：2021年2月9日(周二)20:00

地 点：腾讯会议 ID：309 219 439无密码

摘要：

Machine-learning techniques are efficient tools for discovering structures in high dimensional complex data and have demonstrated their power in many physics disciplines. In this talk, we present a versatile machine-learning method-- tensorial kernel support vector machine (TK-SVM), which can act as an efficient and unsupervised scheme to explore the phase diagram of many-body spin systems and as a comprehensive way to scrutinize spin liquid candidates. The method is illustrated by detecting high-rank tensorial orders emerged from a gauge theory and partitioning the phase diagram of a pyrochlore magnet. It is then applied to Kitaev materials, where the machine identifies several unconventional orders and disordered regimes that are hitherto missed by other approaches. The results in the latter example also bring new insight to the phase diagram of the representative Kitaev material alpha-RuCl3.

简历：

Dr. Ke Liu received his Ph.D. from the University of Leiden under the supervision of Prof. Jan Zaanen in 2016. He now works as a postdoctoral researcher in Prof. Lode Pollet's group at the University of Munich (LMU). His research interests lie in emergent phenomena, phase transitions, and gauge theories in strongly correlated systems. The current focus is to develop new methods to search and understand exotic states of matter, such as tensorial orders and spin liquids, in frustrated magnets.

联系方式： yangfan_blg@bit.edu.cn

邀请人：杨帆 长聘教授

网 址：http:/

承办单位：物理学院、先进光电量子结构设计与测量教育部重点实验室