时间：2017/12/12 (二) 15:30~17:30
地点：政治大学逸仙楼二楼201室 (台北市文山区指南路二段64号)
主题：Robust Optimization, Empirical Likelihood, and Regularization
主办单位：
政治大学商学院风险与保险研究中心
政治大学商学院CARDIF银行保险研究发展中心
政治大学商学院保险业永续发展研究中心
政治大学风险管理与保险学系
财团法人继耘保险文教基金会
| 时间 | 预定会议内容 | |
| 15:30~16:00 | 报到 | |
| 16:00~16:10 | 讲者介绍 | 政治大学商学院风险管理与保险学系 谢明华教授 |
| 16:10~17:10 | 学术专题演讲 |
主讲人： Professor Peter W. Glynn 美国Stanford大学教授 Glynn 博士于 2010 年获得 John von Neumann Theory Prize 的殊荣，并于 2012 年入选为美国国家工程学院院士 (National Academy of Engineering)。目前为 Journal of Applied Probability & Advances in Applied Probability 主编，研究专长为 Simulation, Computational Probability, Statistical Inference for Stochastic Processes, 与 Stochastic Modeling. 演讲主题： Robust Optimization, Empirical Likelihood, and Regularization |
| 17:10~17:25 | 问题与讨论 Q & A | |
| 17:25~17:30 | 闭幕致词 | 政治大学商学院 郑宗记副院长 |
Robust Optimization, Empirical Likelihood, and Regularization
摘要：
We discuss stochastic optimization problems, in which the underlying probability distribution is unknown and needs to be estimated from observed data. Such optimization problems arise in many different settings, including online applications, finance, etc. We make clear the connection between a certain class of robust optimization procedures and generalized empirical likelihood. From this connection, we show how one can construct confidence intervals for optimal values and solutions that achieve exact coverage asymptotically. A key finding is that the distributionally robust formulation essentially regularizes the problem via the sample variance. This work is joint with John Duchi and Hongseok Namkoong.
报名网址：https://goo.gl/hv6noP
场地名额有限，请尽早报名，以免向隅！
承办人：政大风险与保险研究中心 林冠杏小姐、张维娟小姐
电话：(02)2939-3091转65331/65307
email：angellin@nccu.edu.tw
