报告题目:
Spectral detection of global structures in the data by learning a regularization
报告时间:2017年8月10日 ()下午4:00
报告地点:格致楼5004报告厅
报告摘要:Spectral methods are population in detecting global structures in the data. However they have localization problem when data is sparse or noisy. In this talk Pan Zhang will introduce a learning-based method for solving the localization problem. His method is inspired by localization of wavefunctions and its perturbation analysis.
报告人简介:张潘,中科院理论物理研究所副研究员。本科、博士毕业于bet356官网在线登录bet356亚洲版体育官网,后分别在意大利都灵理工,法国ESPCI和美国Santa Fe Institute做博士后研究。现研究方向为统计物理与统计推断、机器学习的交叉领域,目前关心自旋玻璃理论和消息传递算法,网络中的聚类与推断问题,以及无监督神经网络的学习问题。