Recently, the China Association for Science and Technology （CAST） released the result of the 8th Outstanding Scientific and Technological Papers Selection Program, and the paper Machine learning the nuclear mass by Prof. Li Qingfeng and Prof. Wang Yongjia from Huzhou University was on the list. This is the first paper of the University to be selected, and is also the first paper related to nuclear physics theory among all the outstanding scientific and technological papers.

　　This research team is one of the first teams in China to use machine learning methods to study nuclear physics issues. In recent years, the team has taken full advantages of these methods and made important breakthroughs in such frontiers of nuclear physics as nuclear mass, nuclear matter equation of state, density-dependent symmetry energy, fusion reactions, collision parameters, etc., providing a new paradigm for the study of related nuclear physic issues.

　　Atomic mass is the most basic property of atomic nuclei, which is of great significance to understand fundamental interactions and the origin of elements. However, due to the complexity of the nucleus, there are still difficulties and challenges in the theoretical prediction of the nucleus mass. The team of Prof. Li applied machine learning to the research of nuclear physics issues, and demonstrated strong ability of machine learning in exploring key information from nuclear physics related data through detailed analysis of the generalization ability and interpretability of machine learning models. By using LightGBM in machine learning, the team optimized the nuclear mass model by learning the residuals of the theoretical and experimental values, and achieved nuclear mass prediction results of the highest precision at that time. At the same time, the team also carried out detailed discussion on a series of hot issues, including the extrapolation ability, interpretability analysis, uncertainty quantification, exploration of new magic number, neutron separation energy and neutron drip-line positions.

　　The paper was published in the Nuclear Science and Techniques （Q1 in Science and Technology, Volume 32, Issue 10, cover paper） included in SCI, and has gained wide attention in China and beyond. According to Google Scholar, the paper has been cited more than 50 times in just two years, including Peking University, Tsinghua University, Fudan University, University of Chinese Academy of Sciences, MIT, Michigan State University, Jefferson Lab and other academic institutions, as well as research teams from Italy, Colombia, India, Turkey and other countries. Other papers with reference to this article are published in the Reviews of Modern Physics, Science China Physics, Mechanics & Astronomy, Physics Letters B, Physical Review C, Nuclear Physics A, Journal of Physics G: Nuclear and Particle Physics, Chinese Physics C, Nuclear Science and Techniques and other high-level professional journals in the field.

　　The 8th Outstanding Scientific and Technological Papers Selection Program initiated by CAST aims to follow the important instructions of President Xi Jinping to build first-class academic journals and encourage sci-tech workers to publish more high-level research papers in domestic journals. The purpose is to promote high-quality development of Chinese scientific and technological journals internally, and lay solid foundation for sci-tech self-reliance and self-strengthening at higher levels. The whole process of selection involves recommendation from experts in various disciplines and fields, first-round selection, final evaluation, and publication to society, with a total of 201 papers in 81 disciplines selected.

　　Paper link: https://link.springer.com/article/10.1007/s41365-021-00956-1