RRAM Crossbar-Based Fault-Tolerant Binary Neural Networks (BNNs)

Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

12 Downloads (Pure)

Abstract

Computation-In Memory (CIM) using RRAM crossbar array is a promising solution to realize energy-efficient neuromorphic hardware, such as Binary Neural Networks (BNNs). However, RRAM faults restrict the applicability of CIM for BNN implementation. To address this issue, we propose a fault tolerance framework to mitigate the impact of RRAM faults on the accuracy of CIM-based BNN hardware. Evaluation results using MNIST, Fashion-MNIST and CIFAR-10 datasets demonstrate that the proposed framework outperforms the related works as it restores more than 99% of the RRAM fault induced accuracy reduction with relatively less overhead.
Original languageEnglish
Title of host publicationProceedings of the 2022 IEEE European Test Symposium (ETS)
Place of PublicationDanvers
PublisherIEEE
Pages1-2
Number of pages2
ISBN (Electronic)978-1-6654-6706-3
ISBN (Print)978-1-6654-6707-0
DOIs
Publication statusPublished - 2022
Event2022 IEEE European Test Symposium (ETS) - Barcelona, Spain
Duration: 23 May 202227 May 2022

Conference

Conference2022 IEEE European Test Symposium (ETS)
Country/TerritorySpain
CityBarcelona
Period23/05/2227/05/22

Bibliographical note

Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

  • CIM
  • fault tolerance
  • RRAM
  • BNN

Fingerprint

Dive into the research topics of 'RRAM Crossbar-Based Fault-Tolerant Binary Neural Networks (BNNs)'. Together they form a unique fingerprint.

Cite this