[12p-W933-1~11] 31.1 Focused Session "AI Electornics"

△ [12p-W933-8] Demonstration of reservoir computing by using a random network consisting of carbon nanotube complexed with polyoxiometarate molecule

〇Yuki Takeshima¹, Kian Lian Goh¹, Masaru Okada¹, Yuji Kuwahara¹, Tetsuya Asai², Megumi Akai^1,3 (1.Osaka Univ., 2.Hokkaido Univ., 3.JST PRESTO)

Keywords:reservoir computing

We constructed a CNT random network complexed with polyoxiometarate(POM) molecule to perform a reservoir computation. The output current response against an input voltage in the POM/CNT network shows rich dynamics and enough diversity depending on each different point. To produce such rich dynamics in the response, CNTs should be spaced sparsely and appropriately. We successfully demonstrated a reservoir computation by using 50 output signals from a POM/CNT network to learn a cubed sinusoidal signal under a single sinusoidal input.