Neural Network Design via Voltage-based Resistive Processing Unit and Diode Activation Function - A New Architecture

Yung Ting Hsieh, Khizar Anjum, Songjun Huang, Indraneel Kulkarni, Dario Pompili

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Scopus citations

Abstract

In recent years, the architecture based on Resistive Processing Unit (RPU) has become a hot topic due to its potential to accelerate training of a Neural Network (NN). However, attempts to realize the RPU concept based on non-volatile memory technology face a myriad of technological and physical constraints. The theoretical concept of crossbar array is nearly impossible to implement in the real world without certain tweaks. Hence, we propose an Voltage output Complementary Metal Oxide Semiconductor (CMOS)-based RPU design VRPU, which is used to build a neural network. We also introduce a novel diode-based circuit to behave as a non-linear activation function, which consists of a single Diode (D) and Diode Pair (DP). The proposed VRPU design when tested with MNIST dataset for hidden layer and output layer combinations of ReLU+Sigmoid, D+Sigmoid, ReLU+DP, D+DP (low temperature) and D+DP (high temperature) resulted in accuracies of 94.29%, 95.90%, 95.53%, 96.75% and 96.57% respectively corroborating the merits of the proposed design.

Original languageEnglish (US)
Title of host publication2021 IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2021 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages59-62
Number of pages4
ISBN (Electronic)9781665424615
DOIs
StatePublished - Aug 9 2021
Event2021 IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2021 - Virtual, East Lansing, United States
Duration: Aug 9 2021Aug 11 2021

Publication series

NameMidwest Symposium on Circuits and Systems
Volume2021-August
ISSN (Print)1548-3746

Conference

Conference2021 IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2021
Country/TerritoryUnited States
CityVirtual, East Lansing
Period8/9/218/11/21

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Electrical and Electronic Engineering

Keywords

  • Diode-based Activation Function
  • Memristor Crossbar Array
  • Neural Network
  • Resistive Processing Unit

Fingerprint

Dive into the research topics of 'Neural Network Design via Voltage-based Resistive Processing Unit and Diode Activation Function - A New Architecture'. Together they form a unique fingerprint.

Cite this