The Advanced Data Science Project (ADSP) conducts developmental research on data science in the healthand medical elds and on designing methodologies to solve social issues via data science.

Data-driven ｍedical science has been advanced as supervised learning to combine prior knowledge intolearning algorithms, divided into two major paradigms of AI, "symbolic processing and pattern recognition".This project will develop the Platform (PF) in Phase I, by collaborating with Data Folders (DHs) of universityhospitals and other institutions. The next trend of data-driven ｍedical science will be the integration of thetwo paradigms (horizontal integration Phase II) and the integration of cyber (symbolic processing and pattern recognition) andphysical (real-world problems) (vertical integration Phase III). Toward of the realization of Phase II and Phase III, we will developa highly versatile and original middleware-based PF with built-in reliability guarantees for inference.