Data-driven Materials Design Group

There is a need to search for new materials that are truly useful using data-driven methods. By integrating data science, computational science, theoretical science, and experimental data from inside and outside NIMS, this research group will establish a new mechanism for materials design, and synthesize and evaluate new materials in collaboration with experimental researchers. To this end, we will develop materials informatics methods, accumulate computational data, and accumulate know-how on the use of each material system.

Field of Expertise/Research Subjects

In order to use Data-driven Materials Design Group to design new materials, the group is conducting the following research

• Exploration of elemental composition and experimental conditions of materials through active learning
• Extraction and design of feature information required for data analysis
• Development of dimensional compression and structure extraction methods for spectral and image data, and automatic data creation methods
• Reaction mechanism analysis and feature extraction at solid-liquid interface by first-principles molecular dynamics calculation and classical MD simulation
• Efficient accumulation of features and experimental data through autonomous, automated, high-throughput experiments
• Development of automated molecule generation methods for extrapolation search
• Development of material search method for small data