报告题目：Personalized Federated Learning for Sparse High-dimensional Quantile Regression

报 告 人：孔令臣教授（北京交通大学）

报告时间：2025年9月26日下午14:30-15:30

报告地点：龙洞校区行政楼610

Abstract: Multi-source data analysis is an active research topic in modern statistical and data sciences. Its purpose is to obtain accurate estimates by mining multiple sets of data. This field covers a range of approaches such as federated learning, transfer learning, and fusion learning. This paper considers the high-dimensional regularized quantile regression models with data heterogeneity and proposes an alternating direction method of multipliers based robust personalized federated learning algorithm. Rigorous and numerical convergence of the algorithm and the statistical properties of the estimators are established. Simulations and real data analysis confirm the effectiveness of the proposed method.

专家简介：孔令臣，北京交通大学教授，博士生导师，中国运筹学会数学规划分会理事长，北京交通大学数学与统计学院副院长。主要从事对称锥互补问题和最优化、高维数据分析、统计优化与学习、医学成像等方面的研究。在《Mathematical Programming》、《SIAM Journal on  Optimization》、《Statistica Sinica》、《IEEE Transactions on Pattern Analysis and Machine Intelligence》、《Technometrics》、《IEEE Transactions on Signal Processing》和《Electronic Journal of Statistics》等期刊发表论文60余篇。2005年获山东省高等教育教学成果奖三等奖， 2012年获中国运筹学会青年奖，2018年获得北京市高等教育教学成果奖一等奖，2022年获教育部自然科学奖二等奖和北京市高等教育教学成果奖二等奖。