Mitigate IR-Drop Effect by Modulating Neuron Activation Functions for Implementing Neural Networks on Memristor Crossbar Arrays

The line resistance (LR) in a large-scale memristor crossbar array can cause serious IR-drop problem, degrading the hardware deployment capability of neural networks (NNs). In this work, two innovation schemes from the level of software are proposed to mitigate the hardware IR-drop problem by intent...

Full description

Saved in:
Bibliographic Details
Published in:IEEE electron device letters 2023-08, Vol.44 (8), p.1-1
Main Authors: Song, Danzhe, Yang, Fan, Wang, Chengxu, Li, Nan, Jiang, Pinfeng, Gao, Bin, Miao, Xiangshui, Wang, Xingsheng
Format: Article
Language:English
Subjects:
activation function
Arrays
Artificial neural networks
Biological neural networks
Electric potential
Electrodes
Hardware
IR-Drop
line resistance
Memristor array
Memristors
neural network
Neural networks
Neurons
Resistance
Training
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:The line resistance (LR) in a large-scale memristor crossbar array can cause serious IR-drop problem, degrading the hardware deployment capability of neural networks (NNs). In this work, two innovation schemes from the level of software are proposed to mitigate the hardware IR-drop problem by intentionally modulating the NN activation function before deploying. The methods are evaluated over typical activation functions and various line resistances on MLP and LeNet-5 for MNIST recognition. Results show the methods can significantly improve the tolerance of NNs to IR-drop and recover the accuracy in some extent. The methods require no extra hardware overhead and reduce the complexity of peripheral circuits, which make them more achievable and attractive.
ISSN:0741-3106
1558-0563
DOI:10.1109/LED.2023.3285916