Skip to main navigation Skip to search Skip to main content

A Current-link CMOS Analog Neuron with Simplified Synapse Using a Merged Switch Array

  • Hee Yeon Yang
  • , Yu Jin Noh
  • , Sung Ho Lee
  • , Geun Won Kim
  • , Hang Geun Jeong
  • , Donggu Im*
  • *Corresponding author for this work
  • Jeonbuk National University

Research output: Contribution to conferenceConference paperpeer-review

Abstract

A new current-mode analog neuron with a current link between the soma and the synapse is proposed. The proposed neuron requires fewer number of transistors for its synapse due to the merged switch array used for binary weighting. Thus, the burden of hardware complexity for the current-link current-mode neuron is significantly alleviated. The proposed neuron and SRAM cells for weight storage were designed in 0.18μm CMOS process for verification purposes. The simulated linearity of the neuron is sufficiently good to do without calibration of each neuron. Test chips are under fabrication.

Original languageEnglish
Title of host publicationProceedings - International SoC Design Conference, ISOCC 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages77-78
Number of pages2
ISBN (Electronic)9781728183312
DOIs
StatePublished - 2020.10.21
Event17th International System-on-Chip Design Conference, ISOCC 2020 - Yeosu, Korea, Republic of
Duration: 2020.10.212020.10.24

Publication series

NameProceedings - International SoC Design Conference, ISOCC 2020

Conference

Conference17th International System-on-Chip Design Conference, ISOCC 2020
Country/TerritoryKorea, Republic of
CityYeosu
Period20.10.2120.10.24

Keywords

  • Analog Neurons
  • CMOS
  • Current-mode

Quacquarelli Symonds(QS) Subject Topics

  • Computer Science & Information Systems
  • Engineering - Electrical & Electronic
  • Engineering - Petroleum
  • Data Science
  • Physics & Astronomy

Fingerprint

Dive into the research topics of 'A Current-link CMOS Analog Neuron with Simplified Synapse Using a Merged Switch Array'. Together they form a unique fingerprint.

Cite this