Selective predictive inference
报告人:邹长亮(南开大学)
时间:2024-05-23 14:00-15:00
地点:智华楼四元厅-225
Abstract:
Conformal inference is a popular tool for conducting distribution-free predictive inference in regression. We consider here the scenario of selective conformal inference. Two main problems will be discussed: 1) how to use conformal inference to select specific individuals of interest, and 2) how to construct post-selection prediction intervals. Additionally, we explore the development of algorithms tailored for online settings.
About the Speaker:
邹长亮,南开大学统计与数据科学学院教授。主要从事统计学及其与数据科学领域的交叉研究和实际应用。研究兴趣包括:高维数据统计推断、变点和异常点检测、预测性推断等,在统计学和机器学习领域的权威杂志/会议Ann.Stat.、Biometrika、J.Am.Stat.Asso.、J.Mach.Learn. Res.、Math.Program.、ICML/NIPS/AAAI等上发表论文多篇,入选爱思唯尔“中国高被引学者”。主持基金委优青、杰青、重点项目、重大项目课题和科技部重点研发计划课题等。任教育部科技委委员、全国应用统计专业硕士教学指导委员会委员、中国现场统计研究会副理事长等。