Print Email Facebook Twitter RRAM Crossbar-Based Fault-Tolerant Binary Neural Networks (BNNs) Title RRAM Crossbar-Based Fault-Tolerant Binary Neural Networks (BNNs) Author Gebregiorgis, A.B. (TU Delft Computer Engineering) Zografou, Artemis (Student TU Delft) Hamdioui, S. (TU Delft Quantum & Computer Engineering) Department Quantum & Computer Engineering Date 2022 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. Subject CIMfault toleranceRRAMBNN To reference this document use: http://resolver.tudelft.nl/uuid:b3f526c0-6453-4222-8a96-7280e9dde55a DOI https://doi.org/10.1109/ETS54262.2022.9810414 Publisher IEEE, Danvers Embargo date 2023-07-01 ISBN 978-1-6654-6707-0 Source Proceedings of the 2022 IEEE European Test Symposium (ETS) Event 2022 IEEE European Test Symposium (ETS), 2022-05-23 → 2022-05-27, Barcelona, Spain 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. Part of collection Institutional Repository Document type conference paper Rights © 2022 A.B. Gebregiorgis, Artemis Zografou, S. Hamdioui Files PDF RRAM_Crossbar_Based_Fault ... s_BNNs.pdf 1.3 MB Close viewer /islandora/object/uuid:b3f526c0-6453-4222-8a96-7280e9dde55a/datastream/OBJ/view