材料基因工程加速新材料设计与研发

在未知材料化学成分和性能关系的情况下，通过传统的“试错-纠错”方法研发具有特定功能的新材料成本高且经常失败。随着人工智能和数据驱动的第四科学范式的发展，材料基因工程(MGE)已经成为材料设计与研发的新模式。综述了材料基因工程中高通量计算、材料数据库和人工智能方法的研究进展。介绍了材料高通量计算常用的框架和方法; 阐述了材料数据库在材料数据类型和数据标准两方面的发展现状和有待解决的难题; 总结了人工智能方法在材料关键基础问题中的应用。从高通量可视化计算方法、材料多类型数据库和可视化机器学习框架三方面重点证述了自主开发的多尺度集成可视化的高通量自动计算和数据管理智能平台ALKEMIE。展望了材料基因工程未来的发展趋势。 

Novel material design and development accelerated by materials genome engineering

Without knowing the relationship between the chemical composition and properties of the material, the development of new materials with desired properties based on conventional trial-and-error methods is cost inefficient and sometimes ends fruitlessly. With the development of artificial intelligence and the fourth science paradigm, data-driven science, materials genetic engineering (MGE) has become a new approach for designing novel materials. In this paper, recent advances in high-throughput computation, materials database, and artificial intelligence methods in MGE are reviewed. First, the framework and software for high-throughput computations of materials are introduced. Then, the progress and critical problems of materials databases are presented in terms of types and standard interfaces of materials data. Next, we summarize the applications of artificial intelligence methods in critical issues of materials science. In particular, from the aspects of visualized high-throughput calculation methods, multi-type materials database and visualized machine learning framework, we emphasize an in-house developed platform ALKMIE, featured by multi-scale integration of automatic high-throughput calculation, visualization, and intelligent data management. Finally, we highlight the future directions of MGE.