Summary
International Symposium on Nonlinear Theory and Its Applications
2016
Session Number:A2L-D
Session:
Number:A2L-D-2
Letter Reprodution Simulator for Hardware Design of Cellular Neural Network Consisting of Neurons Using Large-Scale Integration Chip and Synapses Using on-Deposited Amorphous in-Ga-Zn-O Films
Tomoya Kameda, Mutsumi Kimura, Yasuhiko Nakashima,
pp.-
Publication Date:2016/11/27
Online ISSN:2188-5079
Summary:
We are developing hardware of a cellular neural network (CNN) that features low power consumption and high-density integration. In our hardware system, amorphous In-Ga-Zn-O (a-IGZO) films work as synapse. We assume to form the neuron circuits using a large-scale integration chip and synapse devices using a-IGZO films deposited directly on the neuron chip. We modeled a-IGZO films for CNN and implemented it into the simulator to determine the network architecture and device parameters. In this time, we succeeded in confirming that this CNN can learn two letters of Arabic numerals. Moreover, we estimated the time necessary for the learning.