Shenzhen University School of Mathematical Sciences  

Liyuan Scholars Colloquium Session 153  

Lecture Title: Cross-Semantic Transfer Learning for High-Dimensional Linear Regression  

Speaker: Chair Professor Jiang Jiancheng (School of Science, Great Bay University)  

Time: 14:00–15:00, November 30, 2025 (Sunday)  

Location: Classroom 1, Huixing Building, Yuehai Campus, Shenzhen University  

Abstract:  

Existing transfer learning methods for high-dimensional linear regression typically assume a one-to-one correspondence between features in the source and target domains, limiting their applicability to semantically aligned features. However, in many practical applications, while the features in the source and target domains differ, they may share predictive similarities, forming a cross-semantic resemblance. To fully leverage this broader form of transferability, we propose a Cross-Semantic Transfer Learning (CSTL) framework. This method employs a weighted fusion penalty to compare each regression coefficient in the target domain with all coefficients in the source domain, thereby capturing potential cross-semantic associations. The weights are determined by the derivative of the SCAD penalty function, effectively approximating an ideal weighting scheme that preserves transferable signals while filtering out source-specific noise. Computationally, we develop an efficient solution for CSTL based on the Alternating Direction Method of Multipliers (ADMM). Theoretically, we prove that under mild conditions, the CSTL estimator achieves performance comparable to the oracle estimator with high probability. Both simulation studies and real-data analyses demonstrate that CSTL significantly outperforms existing methods in scenarios involving cross-semantic or partial signal similarities.  

Speaker Profile:  

Jiang Jiancheng received his Ph.D. from the Department of Mathematics at Nankai University. In July 2024, he returned to China and was appointed as a Chair Professor at the Great Bay University. Prior to his return, he was a dual-appointed Professor in the Department of Mathematics and Statistics and the School of Data Science at the University of North Carolina at Charlotte, and also served as the Co-Director of the Charlotte Center for Trustworthy AI. His research focuses on core areas of econometrics and statistics, including distributed computing, financial time series, high-dimensional statistical learning, nonparametric smoothing, quantile regression, and artificial intelligence. To date, he has published over 70 peer-reviewed papers in top-tier journals such as Annals of Statistics, JASA, JRSSB, Biometrika, and The Econometrics Journal.  

All faculty and students are welcome!  

School of Mathematical Sciences  

November 20, 2025