Using Information Science for Physics, Creating Information Science from Physics

My research lies at the interface of physics and information science. Modern computational physics faces fundamental limitations in computational resources and simulation time, while information science is confronted with the high power consumption of conventional von Neumann computers. At the same time, physics and information science have continuously driven each other's development, giving rise to new research fields and technologies.

Against this background, I employ machine learning techniques to understand physical phenomena that are difficult to capture through human intuition, while also developing methods to accelerate analyses that are computationally expensive with conventional numerical methods. Moreover, I investigate mathematical models of biological information processing towards physical device implementation, aiming to create new computing technologies by bridging theory, experiments, and hardware. My current research includes machine-learning-based prediction of phase transitions and bifurcations, as well as the construction of mathematical models for low-power computers based on novel computational principles inspired by biological behavior.