Descriptors of material structure are the most important element technologies in materials informatics. Prediction in data science is
based on the similarity of past and future data. For example, two arbitrary materials with similar structure are considered to have
similar properties; this is a principle common to all prediction algorithms. Descriptors, which are tools for measuring proximity, play an important role in determining the success or failure of prediction in data science. However, general-purpose tools for descriptor calculation have not been well developed and this inhibits the promotion of materials informatics in society.

Our group is developing the world's largest comprehensive descriptor library in the field of materials informatics. Target materials take diverse forms in the mathematical representation. We are promoting research and development for a wide variety of objects in materials science, including crystalline materials, electronic states, chemical composition, polymer materials, and seemingly disordered systems such as amorphous materials. We intend to develop an ALL-IN-ONE package that supports all these objects and will thus strengthen the academic and technological foundation of our information-integrated materials development; this is our group's mission definition.