National Institute for Materials Science, Japan
Materials Big Data: Expectation and Reality
Abstract: Big data based artificial intelligence has been greatly expected to change the style of materials research, therefore, improve the efficiency and decrease the cost of materials development. Construction of materials data infrastructure has been highlighted since materials data is the fundamental of this new technology. However, in spite of the great efforts made by researchers and editors all of the world on data collection since 1880’s, and database development since 1990’s, data shortage is still the bottleneck of today’s materials informatics. In this talk, an overview will be given on current status of materials data, including an analysis on the difficulties of data collection and edition. Several example will be introduced to show how to make the best use of the existing data and use machine learning with small data sets. Finally, perspective on breakthrough in materials data and materials informatics will be discussed.
Doctor in materials engineering, Shanghai Institute of Ceramics, Chinese Academy of Science, 1994. Doctor in information science, Nagoya University, 2007. Joined National Institute for Materials Science (NIMS) in 2002, and currently Deputy Director of Research and Services Division of Materials Data and Integrated System. Expertise in materials data and data systems. Databases developed include AtomWork, AtomWork-Adv., CompES-X, etc. Expertise in thermal transport and thermophysical properties from nano- to macro-scale.