TY - JOUR
T1 - Analyzing the Impact of Memristor Variability on Crossbar Implementation of Regression Algorithms With Smart Weight Update Pulsing Techniques
AU - Afshari, Sahra
AU - Musisi-Nkambwe, Mirembe
AU - Sanchez Esqueda, Ivan Sanchez
N1 - Publisher Copyright:
© 2004-2012 IEEE.
PY - 2022/5/1
Y1 - 2022/5/1
N2 - This paper presents an extensive study of linear and logistic regression algorithms implemented with 1T1R memristor crossbars arrays. Using a sophisticated simulation platform that wraps circuit-level simulations of 1T1R crossbars and physics-based models of RRAM (memristors), we elucidate the impact of device variability on algorithm accuracy, convergence rate and precision. Moreover, a smart pulsing strategy is proposed for practical implementation of synaptic weight updates that can accelerate training in real crossbar architectures. Stochastic multi-variable linear regression shows robustness to memristor variability in terms of prediction accuracy but reveals impact on convergence rate and precision. Similarly, the stochastic logistic regression crossbar implementation reveals immunity to memristor variability as determined by negligible effects on image classification accuracy but indicates an impact on training performance manifested as reduced convergence rate and degraded precision.
AB - This paper presents an extensive study of linear and logistic regression algorithms implemented with 1T1R memristor crossbars arrays. Using a sophisticated simulation platform that wraps circuit-level simulations of 1T1R crossbars and physics-based models of RRAM (memristors), we elucidate the impact of device variability on algorithm accuracy, convergence rate and precision. Moreover, a smart pulsing strategy is proposed for practical implementation of synaptic weight updates that can accelerate training in real crossbar architectures. Stochastic multi-variable linear regression shows robustness to memristor variability in terms of prediction accuracy but reveals impact on convergence rate and precision. Similarly, the stochastic logistic regression crossbar implementation reveals immunity to memristor variability as determined by negligible effects on image classification accuracy but indicates an impact on training performance manifested as reduced convergence rate and degraded precision.
KW - RRAM
KW - crossbar array
KW - machine learning
KW - stochastic regression
KW - variability
UR - http://www.scopus.com/inward/record.url?scp=85124109323&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85124109323&partnerID=8YFLogxK
U2 - 10.1109/TCSI.2022.3144240
DO - 10.1109/TCSI.2022.3144240
M3 - Article
AN - SCOPUS:85124109323
SN - 1549-8328
VL - 69
SP - 2025
EP - 2034
JO - IEEE Transactions on Circuits and Systems I: Regular Papers
JF - IEEE Transactions on Circuits and Systems I: Regular Papers
IS - 5
ER -