[14p-K102-1~6] Developments in materials databases: Accumulating, extracting, and overlooking knowledge

4:20 PM - 4:50 PM JST(7:20 AM - 7:50 AM UTC)

[14p-K102-6] Material Classification and Feature Extraction of Promising Material Group based on Machine Learning

〇Akira Takahashi¹ (1.Science Tokyo)

Keywords:

Materials informatics,Machine learning,Interpretable AI

We have developed a method for classifying materials based on machine learning that identifies groups of materials sharing similar target properties as well as common features in their constituent elements and atomic arrangements, with the aim of extracting characteristics of elemental composition and atomic configuration from materials databases to achieve desired properties. In this presentation, we will explain this methodology and introduce examples of its application to large-scale computational materials databases.