[F-6] Charge-Based Memory Devices for AI Applications

[F-6-02] Synaptic Devices Based on 3D-Semicircular NAND Flash Memory

〇Seongbin Oh^1,2, Seungwhan Kim^1,2, Ho-Nam Yoo^1,2, Woo Young Choi^1,2, Jong-Ho Lee^1,2 (1. Seoul National Univ. (Korea), 2. Inter-University Semiconductor Research Center (Korea))

Presentation style: On-site (in-person)

https://doi.org/10.7567/SSDM.2022.F-6-02

A novel 3D NAND flash memory is proposed which
uses a single hole as two strings (1H2S) for low-power and
high-density neuromorphic computing. Independent operation of two strings and selective program/erase operation were verified by using 3D-TCAD simulation. Finally,
the spiking neural networks (SNNs) based on a 3D NAND
synaptic array achieved 88.25% accuracy when applied to
the CIFAR-10 data sets.