• 统计研究中心
当前位置: 首页> 系列讲座> 正文

圣母大学LIU FANG教授: Recent advanaces in differentially private data synthesis, statistical learning, and machine learning

光华讲坛——海外名家讲堂第   期

主题: Recent advanaces in differentially private data synthesis, statistical learning, and machine learning

主讲人: 圣母大学LIU FANG教授

主持人: 统计学院林华珍教授

时间2021年12月15日(周三)上午10:30-11:30

举办地点腾讯会议号333-667-996

主办单位统计研究中心和统计学院 国际交流与合作处 科研处


主讲人简介(须有英文介绍)

Dr. Liu   is Professor and Associate Chair in the department of Applied and   Computational Mathematics and Statistics at the University of Notre Dame,   Indiana, USA. She holds a doctoral degree   in Biostatistics from University of Michigan, Ann Arbor. Dr. Liu's main   research interests include data privacy and differential privacy, Bayesian   Statistics, statistical machine learning, missing data analysis, and   applications of statistics to medical, biological, and social sciences. At   Notre Dame, she is an affiliated faculty member at the Harper Cancer Research   Institute, the Eck Institute of Global Health, the Technology Ethics Center,   and the Lucy Family Institute for Data and Society. Dr. Liu’s research has   been generously supported by the National Science Foundation, the National   Institutes of Health, nonprofit foundations and organiztaion, and Notre Dame   internal grants. Dr. Liu is an elected fellow of the American Statistical   Association.

刘芳,美国印第安纳州圣母大学应用与计算数学与统计学系的教授兼副系主任。她在美国密歇根大学安娜堡分校获得生物统计学博士学位。她的主要研究兴趣包括数据隐私和差分隐私,贝叶斯统计,统计机器学习,缺失数据分析,以及统计学在医学,生物和社会科学中的应用。在圣母大学,她是Harper癌症研究所,全球健康Eck研究所,技术伦理中心和Lucy Family数据与社会研究所的affiliated faculty member。她的研究得到了美国国家科学基金会(NSF)、美国国立卫生研究院(NIH)、非营利基金会和组织以及圣母大学(Notre Dame)资助。她是美国统计协会(ASA)的elected fellow。

内容简介(须有英文介绍)

Differential privacy is a state-of-the-art concept in data privacy reseach and applications. In this talk, I will present some recent methodological and application work my group has developed, including methods for differentially private data synthesis and its applications in tabular data, graphs, and location data; differentially private empirical risk minimization with dual-purpose regularizer; and differentially private machine learning through iterative optimization procedures. Statistical challenges and opportunies will also be discussed.

差分隐私是数据隐私研究和应用程序中最先进的概念。本报告中将介绍该团队最近开发的一些方法论和应用工作,包括差分私有数据合成方法及其在表格数据,图形和位置数据中的应用;使用双重用途正则化器的差异化私人经验风险最小化;以及通过迭代优化过程进行差分私有机器学习。还将讨论统计方面的挑战和机会。


上一条:中国人民大学刘勇副教授:无监督泛化误差分析初探

下一条:宾夕法尼亚州立大学薛凌洲副教授:An additive graphical model for discrete data