EEIS 東京大学大学院 工学系研究科 電気系工学専攻

KOHNO Takashi Professor

Komaba Campus

Complex Systems & Bioelectronics
Life/Health/Medical informatics
Electron device/Electronic equipment

Neuromimetic Systems: mimicing the nervous system to achieve robust and intelligent information processing systems

We study silicon neuronal networks as a foundation for the next-generation brain-compatible AI. Our approach is multidisciplinary over phenomenological, constructive, and theoretical fields. Anyone interested in the electronic circuit design and nervous system modeling are welcome.

Research field 1

Analog Silicon Neuronal Network Circuits

The electrical physiological activity of the nervous system is reproduced with ultra-low power consumption by utilizing the physical characteristics of the transistors. It is also necessary to build information processing models that utilize physical noise, which is equivalent to the nervous system. By utilizing various devices as well as CMOS transistors, we aim to achieve the same intellectual information processing as the brain with the same energy efficiency as the brain.
Research field 2

Digital Silicon Neuronal Circuits

Building digital circuits that efficiently simulate large-scale nervous system models is the scope of this topic. Here, the theory of nonlinear dynamics is utilized to build qualitative nervous system models that can be implemented efficiently in the digital arithmetic circuit. We have realized a compact circuit that can precisely reproduce the complex behavior of neuronal cells. We are also studying technologies to connect multiple integrated circuit chips to achieve larger silicon neuronal networks.
Back to the list