Artificial Intelligence in Medicine

抄録

We have been developing artificial intelligence (AI) systems for medicine. Our university, Tokyo Medical and Dental University, launched AI Hospital Development Unit in our Medical Innovation Consortium in 2018. It might be an advanced organization for the collaboration between medical engineering and clinic. The AI systems we have developed consist of inductive, deductive, and abductive inferences. Inductive inference includes statistics and deep neural networks, which might be the systems capturing frequently-occurred data patterns. Deductive inference has the capability to integrate their tasks with theoretical manners. Abductive inference is needed to integrate the knowledge, which was stored with inductive and deductive inferences, and the measures of patients. Our systems direct to be integrated monolithically and work cooperatively, which we call intelligently orchestrating network of AIs (iON-AI). We will present these AIs and their feasibility with practical evidence.