0


有限实验的因果可转移性

Causal Transportability with Limited Experiments
课程网址: https://videolectures.net/videos/nipsworkshops2013_bareinboim_lim...  
主讲教师: Dengyong Zhou
开课单位: 信息不详。欢迎您在右侧留言补充。
开课时间: 2014-10-06
课程语种: 英语
中文简介:
大多数科学探索都涉及将实证发现推广到新的环境、设置或人群,机器学习文献将这个问题称为“迁移学习”。我们的演讲侧重于一种特殊类型的泛化能力,称为“可迁移性”,定义为将实验研究中学习到的因果效应转移到新人群的许可证,在这种人群中只能进行观察性研究。我们引入了一种称为“选择图”的形式化表示。
课程简介: Most scientific explorations are concerned with generalizing empirical findings to new environments, settings, or populations, a problem that the machine learning literature labeled "transfer learning." Our talk focuses on a particular type of generalizability, called “transportability”, defined as a license to transfer causal effects learned in experimental studies to a new population, in which only observational studies can be conducted. We introduce a formal representation called “selection diagrams”.
关 键 词: 机器学习; 观察性研究; 科学探索
课程来源: videolectures
数据采集: 2025-05-29:yuhongrui
最后编审: 2025-05-29:yuhongrui
阅读次数: 5