[19p-Z34-1~17] FS.1 Focused Session "AI Electronics"

△ [19p-Z34-5] Physical Reservoir Computing on Atomic Switch Arrays with High Precision and Excellent Memory Characteristics

〇(M1)Hiroshi Kubota¹, Tsuyoshi Hasegawa², Megumi Akai-Kasaya¹, Tetsuya Asai¹ (1.Hokkaido Univ., 2.Waseda Univ.)

Keywords:reservoir computing, atomic switch

Reservoir computing (RC) is an artificial neural network architecture that exhibits temporal dynamic behavior, and its computational structure can be implemented in physical systems. We propose RC with atomic switches. Since the resistance of atomic switches changes nonlinearly with a change in input, the RCs are expected to have a high precision and large memory capacity. We simulated this architecture with Python. The results showed that our RC had higher precision and larger memory capacity.