Machine learning strategies for small sample size in materials science

作     者：Qiuling Tao JinXin Yu Xiangyu Mu Xue Jia Rongpei Shi Zhifu Yao Cuiping Wang Haijun Zhang Xingjun Liu 陶秋伶;于金鑫;穆翔宇;贾雪;施荣佩;姚志富;王翠萍;张海军;刘兴军

作者机构：School of Materials Science and Engineeringand Institute of Materials Genome&Big DataHarbin Institute of TechnologyShenzhenShenzhen 518055China School of Computer Science&TechnologyHarbin Institute of TechnologyShenzhenShenzhen 518055China Advanced Institute for Materials Research(WPI-AIMR)Tohoku UniversitySendai 980-8577Japan Department of Materials Science and EngineeringCollege of MaterialsXiamen UniversityXiamen 361005China State Key Laboratory of Advanced Welding and JoiningHarbin Institute of TechnologyShenzhenShenzhen 518055China 

出 版 物：《Science China Materials》 (中国科学(材料科学)(英文版))

年 卷 期：2025年第68卷第2期

页      面：387-405页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0805[工学-材料科学与工程（可授工学、理学学位）] 080502[工学-材料学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by the National Natural Science Foundation of China (52371007 and 52301042) the National Key R&D Program of China (2020YFB0704503) the Guangdong Basic and Applied Basic Research Foundation (2021B1515120071) the Key-Area Research and Development Program of Guangdong Province (2023B0909050001) 

主　　题：material design machine learning small sample size few-shot learning material domain knowledge 

摘      要：Machine learning (ML) has been widely used todesign and develop new materials owing to its low computational cost and powerful predictive capabilities. In recentyears, the shortcomings of ML in materials science have gradually emerged, with a primary concern being the scarcity ofdata. It is challenging to build reliable and accurate ML modelsusing limited data. Moreover, the small sample size problemwill remain long-standing in materials science because of theslow accumulation of material data. Therefore, it is importantto review and categorize strategies for small-sample learningfor the development of ML in materials science. This reviewsystematically sorts the research progress of small-samplelearning strategies in materials science, including ensemblelearning, unsupervised learning, active learning, and transferlearning. The directions for future research are proposed, including few-shot learning, and virtual sample *** importantly, we emphasize the significance of embedding material domain knowledge into ML and elaborate on thebasic idea for implementing this strategy.