Presentation 2022-08-18
Neuromorphic computing using photo-induced current properties in ITO/Nb:SrTiO3 junction
Yutaro Yamazaki, Hiroyuki Kai, Kentaro Kinoshita,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Recently, needs for edge computing have been increasing, and methods to reduce computational cost while maintaining high learning performance have been required. One of the candidates is physical reservoir computing (PRC). We propose a PRC device that utilizes the photo-induced current properties of the ITO/Nb:SrTiO$_3$ junction as a reservoir. In this study, we show that the transient photo-induced current properties of the ITO/Nb:SrTiO$_3$ junction can be electrically controlled. To evaluate the device's PRC performance, we converted image data into pseudo-time series signals and performed a classification task, which revealed that the learning was optimized by matching the timescale of the device's transient response to that of the input signal. These results indicate the possibility of using the ITO/Nb:SrTiO$_{3}$ junction as a PRC device to realize an optical learning edge AI device that can learn a wide range of signals with a single device.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) reservoir computing / neuromorphic computing / SrTiO3 / artificial optoelectronic synapse / resistive random access memory / memristor
Paper # ED2022-21
Date of Issue 2022-08-11 (ED)

Conference Information
Committee ED / IEE-BMS / IEE-MSS
Conference Date 2022/8/18(1days)
Place (in Japanese) (See Japanese page)
Place (in English) Kikai-Shinko-Kaikan Bldg. B3-2
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Hiroki Fujishiro(Tokyo Univ. of Science) / 工藤 寛之(明治大学) / 佐々木 実(豊田工業大学)
Vice Chair Seiya Sakai(Hokkaido Univ.) / / 寒川 雅之(新潟大学)
Secretary Seiya Sakai(Saga Univ.) / (NTT) / 寒川 雅之(豊橋技術科学大学)
Assistant Masatoshi Koyama(Osaka Inst. of Tech.) / Yoshitugu Yamamoto(Mitsubishi Electric) / / 神田 健介(兵庫県立大学)

Paper Information
Registration To Technical Committee on Electron Devices / Technical Committee on Bio Micro Systems / Technical Committee on Micromachine and Sensor System
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Neuromorphic computing using photo-induced current properties in ITO/Nb:SrTiO3 junction
Sub Title (in English) For reservoir computing application
Keyword(1) reservoir computing
Keyword(2) neuromorphic computing
Keyword(3) SrTiO3
Keyword(4) artificial optoelectronic synapse
Keyword(5) resistive random access memory
Keyword(6) memristor
1st Author's Name Yutaro Yamazaki
1st Author's Affiliation Tokyo University of Science(Tokyo Univ. of Sci.)
2nd Author's Name Hiroyuki Kai
2nd Author's Affiliation Tokyo University of Science(Tokyo Univ. of Sci.)
3rd Author's Name Kentaro Kinoshita
3rd Author's Affiliation Tokyo University of Science(Tokyo Univ. of Sci.)
Date 2022-08-18
Paper # ED2022-21
Volume (vol) vol.122
Number (no) ED-152
Page pp.pp.17-20(ED),
#Pages 4
Date of Issue 2022-08-11 (ED)