活动时间

    2026年9月9日（周三）15:00

    活动地点

    中国人民大学通州校区（具体地点将在报名后于群内通知）

    沙龙主题

    Similarity-Informed Transfer Learning for Post–kidney Transplant Survival Time

    活动内容

    To address the challenge of estimating post–kidney transplant survival time based on functional clinical biomarkers under limited sample sizes and strict data-sharing privacy constraints, we propose a Similarity-Informed Transfer Learning (SITL) method that enables reliable estimation and inference. SITL selectively borrows information from source cohorts by assigning weights based on their similarity to the target cohort and their sample sizes. This weighted borrowing effectively reduces the complexity of the target estimation task and improves estimation efficiency. Rather than relying solely on small local data or indiscriminately pooling heterogeneous sources, SITL enables more stable and accurate estimations of functional effects. Building upon SITL, we further develop a resampling based inference procedure that constructs confidence intervals for functional coefficients while accounting for the transferred information with theoretical guarantees. In the kidney transplant application, SITL showcases substantial improvements in estimation and inference over methods that use only the target cohort or naively pool all data. These improvements are also confirmed through extensive simulation studies. Such advantages are particularly critical given the common constraints of limited sample sizes and stringent privacy requirements in medical settings. Overall, this method provides a general framework for applications with limited samples, particularly in biomedical research, offering guidance for reliable estimation and inference.

    邀请嘉宾

    曹际国

    嘉宾简介

    曹际国博士, 加拿大温哥华西蒙弗雷泽大学(Simon Fraser University)统计与精算系教授，加拿大数据科学国家特聘教授(Canada Research Chair in Data Science)，现担任Statistics in Medicine等国际统计期刊副主编。曹际国教授2006年获得加拿大麦吉尔大学(McGill University) 博士，2007年美国耶鲁大学博士后出站，长期从事人工智能，机器学习，函数型数据分析(functional data analysis) 和估计微分方程的研究。曹际国教授2021年获得加拿大统计协会(Statistical Society of Canada)和国家数学研究中心(Centre de recherches mathématiques)联合评比的最高奖之一：加拿大国家杰出青年统计学家奖(CRM-SSC award)。曹际国教授近些年来在国际优秀期刊中发表超过150篇文章。

    报名方式

    如您有意参加，请于2026年9月9日（周三）之前扫描下方问卷星二维码报名并入群，会议具体地点将在群内通知，感谢您的关注与支持！