In order to perform reliable simulations on very large systems, such as nano-structured materials, nano-scaled catalysts, and bio-systems, NIMS-CMSU has been developing a linear-scaling DFT method (http://www.linear-scaling.orgv). With this method, we can employ DFT calculations on systems containing more than tens or hundreds of thousands of atoms.
Materials science researches by machine learning are conducted. Various materials data such as simulation, synthesis, and measurement results are targeted, and new materials informatics techniques are also developed.