[20p-M206-1~11] In-materio AI computing

[20p-M206-10] In-materio reservoir computing and robot application by random network made of nanoparticles with chemical dynamics

〇Hirofumi Tanaka^1,2 (1.LSSE, Kyutech., 2.Neumorph Center, Kyutech)

Keywords:in-materio reservoir, Classification of robot hand grabbed products, Echo state network

Recently, reservoir computing (RC) is much more attracted attention because learning is required only in the output layer of the AI system, and energy consumption and accuracy will be higher than those of general devices. Echo state networks (ESN) with non-linear, memory effects and high harmonic can be used as physical hardware devices expected to be replaced by software systems. In this work, we introduce computing results obtained from the materials we are using as ESN RC devices