The 66th JSAP Spring Meeting, 2019

Presentation information

Oral presentation

31 Focused Session "AI Electronics" » 31.1 Focused Session "AI Electornics"

[11p-W810-1~17] 31.1 Focused Session "AI Electornics"

Mon. Mar 11, 2019 1:15 PM - 6:00 PM W810 (E1001)

Jun-ichi Shirakashi(TUAT), Tsuyoshi Hasegawa(Waseda Univ.)

5:30 PM - 5:45 PM

[11p-W810-16] A Crystalline Oxide Semiconductor-FET/Si-FET Hybrid Structure-based Multiplier for Artificial Neural Network Applications

Yoshiyuki Kurokawa1, Takeshi Aoki1, Munehiro Kozuma1, Hajime Kimura1, Shunpei Yamazaki1 (1.SEL)

Keywords:crystalline oxide semiconductor, neural network, IGZO

The crystalline oxide semiconductor FET (OS-FET), which uses a crystalline oxide semiconductor material as its channel layer, exhibits an extremely low off-state current and enables memory capable of long-term data retention. In addition to processors and FPGAs, analog multipliers that exploit the OS-FET's potential ability to add analog data retention capabilities to memory have been proposed as LSI applications. This work shows the measured basic property of an analog multiplier with an OS-FET/Si-FET structure, and evaluates its applicability to neural networks.