Physical Reservoir Computing Using Nonlinear Heat Conduction

As the demand for machine learning and data analysis increases with the rapid informatization of society, physical reservoir computing (PRC) that performs learning with low computational load by using transient state changes of nonlinear physical phenomena is gaining attention.
In this research, we propose a PRC using heat conduction with temperature dependent thermal conductivity, and aim to improve its performance.

Conference Presentation

Y. Takeda, T. Mizumoto, A. Banerjee, T. Tsuchiya, J. Hirotani, ” Thermal reservoir computing using nonlinear temperature coefficient of thermal conductivity”, The 13th Symposium on Micro-Nano Science and Technology, Tokushima, Japan (November 14-16, 2022),14P5-PN-20.