Optimized Near-Zero Quantization Method for Flexible Memristor Based Neural Network

Due to controllable conductance and non-volatility, flexible memristors are regarded as a key enabler for building artificial neural network (ANN)-based learning algorithms in flexible and wearable systems. However, the existing flexible memristors are suffering from limited number of conductance va...

Full description

Saved in:
Bibliographic Details
Published in:IEEE access 2018-01, Vol.6, p.29320-29331
Main Authors: Xu, Jiawei, Huan, Yuxiang, Yang, Kunlong, Zhan, Yiqiang, Zou, Zhuo, Zheng, Li-Rong
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Aluminum oxide
Artificial neural network
Artificial neural networks
Biological neural networks
Circuits
Dielectrics
flexible memristor
Hysteresis
Learning theory
Machine learning
Measurement
Memristors
near-zero optimizing
Neural networks
Polymer matrix composites
Quantization (signal)
Random errors
Resistance
Single wall carbon nanotubes
system resilience
Volatility
weight quantization
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Due to controllable conductance and non-volatility, flexible memristors are regarded as a key enabler for building artificial neural network (ANN)-based learning algorithms in flexible and wearable systems. However, the existing flexible memristors are suffering from limited number of conductance values, issues limiting large-scale integration, and insufficient accuracy that cannot support accurate computation of ANN. In this paper, solutions are proposed for the three major challenges of the flexible memristor; the feasibility of a three-layer fully connected neural network on MNIST and a 13-layer convolutional neural network (CNN) on CIFAR-10 using the flexible memristor based on single-walled carbon nanotubes network/polymer composite and hydrophilic Al 2 O 3 dielectric are studied. The evaluation result shows that in the fully connected neural network system, it is able to recognize MNIST with an accuracy above 90% after 4-bit quantization, 52.05% decrease in interconnection numbers in the circuit and up to 40% random error introduced, and in the CNN on CIFAR-10, the system can retain an accuracy above 86% with less than 4% accuracy loss after 5-bit quantization, 59.34% decrease in interconnection numbers in the circuit and up to 40% random error injected.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2018.2839106