| Paper Abstract and Keywords |
| Presentation |
2024-11-01 12:10
[Invited Talk]
Development of oxide-based leaky-integrating transistor for spiking neural networks Hisashi Inoue (AIST), Hiroto Tamura (Univ. Tokyo), Ai Kitoh (AIST), Xiangyu Chen, Zolboo Byambadorj (Univ. Tokyo), Takeaki Yajima (Kyushu Univ.), Yasushi Hotta (Univ. Hyogo), Tetsuya Iizuka (Univ. Tokyo), Gouhei Tanaka (Univ. Tokyo/Nagoya Inst. Tech.), Isao Inoue (AIST) MRIS2024-26 CPM2024-55 |
| Abstract |
(in Japanese) |
(See Japanese page) |
| (in English) |
Biomimetic computing aims to realize energy-efficient information processing by mimicking the behavior of biological neural networks. Biological neurons exhibit leaky-integrate-and-fire behavior with relatively long leaky integration time scales of milliseconds to seconds. We report solid-state transistor devices that exhibit leaky integration behavior with such a long timescale by exploiting the drift-diffusion of oxygen vacancy ions in solids. A randomly connected spiking neural network called a reservoir was constructed to confirm its operation. Furthermore, we report a demonstration of handwriting anomaly detection as an example task in reservoir computing by numerically simulating a spiking neural network. |
| Keyword |
(in Japanese) |
(See Japanese page) |
| (in English) |
spiking neural network / leaky integration / reservoir / anomaly detection / / / / |
| Reference Info. |
IEICE Tech. Rep., vol. 124, no. 227, MRIS2024-26, pp. 77-80, Oct. 2024. |
| Paper # |
MRIS2024-26 |
| Date of Issue |
2024-10-24 (MRIS, CPM) |
| ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
| Download PDF |
MRIS2024-26 CPM2024-55 |