School of Mathematical Sciences Academic Report [2023] No. 060

(High-level University Construction Series Report No. 831)

Report title: A Correlation-Ratio Transfer Learning and Variational Stein's Paradox

Speaker: Prof. Lin Lu (Shandong University)

Reporting time: Friday, September 1, 2023 10:00-11:00 a.m

Location:Room514,Huixing Building

Report content: A basic condition for efficient transfer learning is the similarity between a target model and source models. In practice, however, the similarity condition is difficult to meet or is even violated. Instead of the similarity condition, a brand-new strategy, linear correlation-ratio, is introduced in this paper to build an accurate relationship between the models. Such a correlation-ratio is estimable by historical data of a part of variables. Then a correlation-ratio transfer learning likelihood is established based on the correlation-ratio combination. On the practical side, the new framework is applied to some application scenarios, especially the areas of data streams and medical studies. Methodologically, some techniques are suggested for transferring the information from simple source models to a relatively complex target model. Theoretically, some favorable properties, including variance reduction and convergence rate improvement, are achieved, even for the case where the source models are not similar to the target model. All in all, it can be seen from the theories and experimental results that the inference about the target model is significantly improved by the information from similar or dissimilar source models. In other words, a variational Stein's paradox is illustrated in the context of transfer learning.

Biography: Lin Lu is a professor and doctoral supervisor at the Institute of Securities and Finance, Shandong University, a member of the First and Second Steering Committee of Master of Education in Applied Statistics of the Ministry of Education, a member of the Steering Committee of Master of Education in Applied Statistics of Shandong Provincial Department of Education, and a Shandong Provincial Government Councillor. He is engaged in research on big data, high-dimensional statistics, nonparametric and semiparametric statistics, and financial statistics, and has published more than 120 research papers in top international journals of statistics, machine learning and related applied disciplines (including Ann. Statist., JMLR, Science China) and other important journals; several financial strategy reports have been positively approved by the governor of Shandong Province; and he has presided over several projects of the National Natural Science Foundation of China, the National Statistics Center of China, the National Statistical Institute of China, the National Statistical Institute of China. He has presided over a number of projects of National Natural Science Foundation of China (NSFC), National Major Project of Statistical Science Research, Doctoral Special Fund of Ministry of Education (MOE), New Liberal Arts Project of MOE, and Key Project of Natural Science Foundation of Shandong Province, etc.; he has won the first and second prizes of National Outstanding Research Achievements in Statistics issued by the National Bureau of Statistics (NBS), and the first prize of Excellent Teaching Achievements of Shandong Province (both ranked first).

Translated with www.DeepL.com/Translator (free version)

Teachers and students are welcome to participate!

Inviter: Hu Zongliang

Schoolof Mathematical Sciences

August 31, 2023