Presentation Information
[1G01]Bridging Structure and Function of Materials: Extended Free Energy Model
*Masato Kotsugi1 (1. Tokyo University of Science)
We have developed a novel materials informatics framework called the extended free energy model (eX-GL), which bridges microstructures and macroscopic functionalities by integrating physical principles with data science. Based on the concept of free energy as a universal descriptor, eX-GL robustly links structural features with energy landscapes, enabling interpretable AI-driven analysis. In thin film growth studies, persistent homology and PCA revealed branching conditions in dendritic structures of Cu films. In magnetic materials, eX-GL visualized domain wall structures and energy loss mechanisms, uncovering factors like coercivity in grain structures previously hidden from conventional analysis. These achievements demonstrate the broad applicability of eX-GL, suggesting its potential as a universal model for materials function analysis derived from experimental data.