理论物理交流平台系列报告—慕尼黑大学刘科博士

发布日期:2018-11-02

应物理学院钟寅老师邀请,慕尼黑大学刘科博士来我院访问,并做学术报告,欢迎广大师生届时参加!

题目:Identifying Unconventional Spin  Order with Interpretable Machine

时间:20181017日星期三,上午10:30

地点:格致楼3016报告厅

摘要:

Frustrated magnetism is one of  the central topics in modern condensed matter physics, as frustration often acts as a firm source of  unconventional states of matter. Canonical examples include various spin  liquids and spin nematics. However, distinguishing these states and  identfying their characterization are usually challenging. In this  talk, we present a machine learning protocol to detect and discern  general spin-orientation orders. We demonstrate the capacity of the  method by learning multipolar order parameters up to rank-6 tensor, and  by exploring phase diagrams of multiple phases. This method may be used  as an alternative tool in studying phase transitions in frustrated spin  and orbital systems, and may prove useful for identifying novel spin  nematics and ruling out spurious spin liquid  candidates.

报告人简历:

刘科,2016年荷兰莱顿大学博士,导师Jan Zaanen教授,现为德国慕尼黑大学博士后。当前研究方向为自旋系统中的奇异序,机器学习,规范场论在凝聚态物理中的应用。代表工作包括广义液晶序参量的分类及其相变普适性质的研究。