Extended free energy model for bridging data science and materials science

Abstract

We will address how measurement informatics can link real-world material functionality with cyber-world data analysis. We have developed an "extended free energy model" that links micro-structure and macro-function through research on applied materials. This model analyzes free energy, a common concept between materials and information, in cyberspace, and can convert information on the relationship between complex structural changes inside materials and material functions in a white-box manner. This model has succeeded in causal analysis of functional mechanisms, extraction of knowledge unnoticed by humans, and optimization of material functions. In the talk, I will introduce actual examples in magnetic materials and discuss its application to the field of artificial intelligence.