On October 20, 2020, invited by Professor Kang Guozheng, Academician Zhang Tongyi of Shanghai University came to our school and give an academic lecture with the topic of "Machine learning of small data". Professor Kang Guozheng, Li Xiangyu, and other teachers and graduate students attended this lecture.

Professor Kang Guozheng, first expressed a warm welcome to Academician Zhang Tongyi and introduced the basic situation of him to the participants. Zhang Tongyi is an academician of the Chinese Academy of Sciences and the Hong Kong Academy of Engineering Sciences, an expert in materials science and solid mechanics, a promoter of material genomic engineering, material informatics and mechanical informatics in China, the founding president of the Institute of material genomic engineering of Shanghai University, and the first director of the material genome engineering branch of China Materials Society.He has served as Chair Professor of the Hong Kong University of Science and Technology, Fong's Named Professor, Vice-chairman of the International Fracture Society and Vice-chairman of the Far East andOceania Society of Fractures.He has won the Croucher Senior Research Scholar Award of Hong Kong, ASM international Fellow Award of the United States, international fracture science Fellow Award, second prize of National Natural Science Award (twice), youth science and technology award of China Association of science and technology, and science and technology progress award of He Liang He Li Foundation. In recent years, he has vigorously promoted material informatics at home and abroad, and he is the first one to propose the new concept of mechanical informatics; advocating the data-driven new material discovery integrating expert knowledge, and the new concept of combining forward design and reverse design of materials; he called for the development of a new model with data as the center, supporting the research and innovation of new materials upward, accelerating the new model of industrial manufacturing production intelligence and information downward.

In this academic lecture, Academician Zhang Tongyi first introduced the role of machine learning in natural science research in a simple way. He mentioned that the development of natural science is a process in whichhuman beings constantly observe and summarize natural behaviors to obtain knowledge. At the same time, taking the famous Paris law in fracture mechanics as an example, he showed us that many of the most basic and widely used sciences are summarized from data rather than derived from related theories. In the past, human beings used to extract knowledge from natural data, but machine learning can replace human individuals to learn knowledge from a large number of data.Then, he introduced the related work of their research group based on machine learning, introducing the size-related strength analysis of concrete materials, emphasizing the role of domain knowledge in the machine learning of small data, and he mentioned statistical analysis of small data may be able to evaluate theoretical models developed from domain knowledge. Finally, he also gave constructive guidance to young teachers and graduate students in research topic selection.

This lecture allowed teachers and students to understand the role of the currently vigorously developed machine learning technology in natural science research, and expanded the research ideas of teachers and students. The entire academic lecture ended with warm applause.

Fig 1. Academician Zhang Tongyi is giving a report

Fig 2. Students are asking academician Zhang Tongyi for advice