[12p-S101-1~15] FS.1 Focused Session "AI Electronics"

[12p-S101-1] [INVITED] Material Intelligence generated by random network:
In-materio physical reservoir computation devices

〇Hirofumi Tanaka^1,2 (1.Neumorph Center, Kyushu Inst. Tech., 2.Graduate School of LSSE, Kyushu Inst. Tech.)

Keywords:In-materio, reservoir, voice recognition

Reservoir computation, a type of recurrent neural network, is expected to be more energy-efficient and perform faster prediction and classification tasks because they require only learning of output weights from multiple output signals. In this study, we construct a random network using various materials, which we have already confirmed to be effective as an Echo State Network type physical reservoir device. The presentation will introduce the results of computational characteristics when used as a physical reservoir.