主题：Rank-based dependence and conditional dependence measure
主讲人：中国科学院数学与系统科学研究院李启寨教授
主持人：统计与数据科学学院周岭教授
时间：2026年11月7日（周六）上午11:00-12:00
地点：柳林校区弘远楼408会议室
主办单位：统计与数据科学学院和统计研究中心 科研处
主讲人简介：
李启寨，中国科学院数学与系统科学研究院研究员，2001年本科毕业于中国科学技术大学，2006年博士毕业于中国科学院数学与系统科学研究院，2006-2009年在美国国立卫生健康研究院国家癌症研究所从事博士后研究；研究方向：生物医学统计、遗传统计、复杂数据推断；在Nature Genetics, Science Advances, Angewandte Chemie International Edition, Cancer Research, Microboiome, American Journal of Human Genetics, Bioinformatics, IEEE Transactions on Pattern Analysis and Machine Intelligence，Psychometrika, JASA, JRSSB等期刊发表学术论文130余篇；曾获美国统计学会会士(ASA Fellow, 2020),国际统计学会当选会员(ISI Elected Member, 2016),中国工业与应用数学学会优秀青年学者奖(2015);主持及曾主持国家自然科学基金5项；现任中国现场统计研究会党委书记、副理事长，中国数学会常务理事等。
内容提要：
Inferring the strength of dependence and conditional dependence and testing them are fundamental problems in statistics. A recent breakthrough by Chatterjee (2021), and Azadkia and Chatterjee (2021) introduced, for the first time, the dependence and conditional dependence measure that equals 0 if and only if the variables under study are independent and conditionally independent, and equals 1 if and only if they are perfectly dependent.
They further proposed a computationally efficient and strongly consistent estimator, based on an ingenious use of ranks and nearest neighbors. Despite these attractive features, the asymptotic theory of Tn has remained largely undeveloped. We proposed a family of Chatterhee's correlation. We also prove that, under general dependence, Tn is asymptotically normal and its limiting variance admits a closed form. We also construct consistent variance estimators that are computationally efficient and implementable in O(nlogn) time. Taken together with existing bias-correction methods, these results provide a complete inferential theory for Tn.
