January 16, 2024
Keio University School of Medicine
RIKEN
Professor Kazuhiro Sakurada of the Ishii-Ishibashi Memorial Course (Extended Intelligence Medicine) at the Keio University School of Medicine (also Team Leader of the Open-ended Information Science Team, Center for Advanced Intelligence Project, RIKEN) and Associate Professor Tetsuro Ishikawa (also Visiting Principal Investigator of the Medical Data Mathematical Reasoning Team in the same project at RIKEN) have successfully developed a physics theory that will serve as the foundation for predictive medical sciences.
The results of this research provide a foundational theory for realizing hybrid AI, which combines physics and artificial intelligence (AI) analysis, and are expected to contribute widely to medical sciences based on high-precision prediction.
Developing technology to individually predict future changes in the characteristics of humans, organisms, societies, and ecosystems, including the onset of diseases, and to prevent problems from occurring is a pressing issue for achieving a safe and secure society. With the advancement of AI technology, surrogate models capable of high-precision prediction from large amounts of real-world data have also been developed in the field of medical sciences. However, surrogate models have a "black box" structure whose operating principles are unknown, which poses a problem: the reproducibility and reliability of predictions can be affected by biases in the data used to build the model. In clinical settings where risk is unacceptable, surrogate models have not yet been fully accepted. To solve this problem, Professor Kazuhiro Sakurada and Associate Professor Tetsuro Ishikawa aimed to create a hybrid AI that combines natural principles with AI analysis, and they developed a pioneering theory for modeling the dynamics of organisms based on the principles of physics acting on gene products.
The results of this research were published online in the international scientific journal Scientific Reports on January 10, 2024.
Please see below for the full press release.