Scientists from Tohoku University and Fujitsu have employed artificial intelligence (AI) to analyze data collected during the study of a superconducting material – cesium vanadium antimonide (CsV3Sb5). This innovative approach has revealed that the superconductivity mechanism in this material is driven by the interaction of electrons from vanadium, antimony, and cesium. The technology developed, based on Fujitsu’s Kozuchi platform, has significantly simplified the analysis of large volumes of data obtained through the method of angle-resolved photoemission spectroscopy (ARPES), used for examining the electronic state of materials.
The model automatically identifies cause-and-effect relationships, allowing researchers to efficiently extract useful information without relying on intuition.
The use of AI has reduced the scale of the causality graph by more than 20 times compared to traditional methods. This has significantly accelerated the process of data analysis and the identification of key factors affecting superconductivity.
The authors believe that the developed technology will facilitate the creation of new functional materials, including high-temperature superconductors and energy-efficient devices, which will help solve global environmental problems. With the growing interest in sustainable technologies, advancements like this could have profound implications for diverse industries, enhancing energy efficiency in electronics and transportation.
Fujitsu plans to provide a trial version of the Kozuchi AI platform to other researchers in March 2026. Recent updates suggest that this move aims to encourage collaborative research across universities and private sector companies globally, extending its impact beyond Japan, with potential integrations in industrial applications.
