Academics

PEARL: Performance-enhanced Aggregated Representation Learning

Time:Thur., 16:00-17:00, Mar. 27, 2025

Venue:B725, Shuangqing Complex Building A

Speaker:Xinyu Zhang

Speaker:

Xinyu Zhang (CAS)

Time:

Thur., 16:00-17:00, Mar. 27, 2025



Venue:

B725, Shuangqing Complex Building A



Abstract:

This paper studies the general framework of representation learning and develops a frequentist model averaging framework to combine different representation learning methods. The weight choice criterion is K-fold cross-validation criterion. We prove the asymptotic optimality and the weight consistency. Numerical studies and real-data analysis demonstrate the superior performance of the proposed method.

DATEMarch 26, 2025
SHARE
Related News
    • 0

      Personalized Bayesian Federated Learning with Wasserstein Barycenter Aggregation

      Yifan Sun 孙怡帆中国人民大学孙怡帆,中国人民大学统计学院教授,博士生导师,人文社科部副部长、数理统计系系主任,教育部人文社会科学重点研究基地应用统计研究中心研究员,全国工业统计学教学研究会常务理事、中国统计教育学会理事,目前主要从事联邦学习和隐私计算领域研究,在NIPS、ICML、AAAI等高人工智能会议和期刊发表学术论文40余篇,主持国家自然科学基金,获教学标兵、北京市高等教育教学成果一等奖等教学奖励。# O...

    • 1

      Model Selection for Optimal Regression Learning

      In statistical learning, various mathematical optimalities are used to characterize performances of different learning methods. They include minimax optimality from a worst-case standpoint and asymptotic efficiency from a rosy view that the regression function to be learned sits there to be discovered. When multiple models, e.g., trees, neural networks and support vector machines, are considere...