{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,6,18]],"date-time":"2025-06-18T04:26:12Z","timestamp":1750220772099,"version":"3.41.0"},"reference-count":37,"publisher":"Association for Computing Machinery (ACM)","issue":"4","license":[{"start":{"date-parts":[[2020,8,18]],"date-time":"2020-08-18T00:00:00Z","timestamp":1597708800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["J. Emerg. Technol. Comput. Syst."],"published-print":{"date-parts":[[2020,10,31]]},"abstract":"<jats:p>Recent advances in deep neural network demand more than millions of parameters to handle and mandate the high-performance computing resources with improved efficiency. The cross-bar array architecture has been considered as one of the promising deep learning architectures that shows a significant computing gain over the conventional processors. To investigate the feasibility of the architecture, we examine non-idealities and their impact on the performance. Specifically, we study the impact of failed cells due to the initialization process of the resistive memory-based cross-bar array. Unlike the conventional memory array, individual memory elements cannot be rerouted and, thus, may have a critical impact on model accuracy. We categorize the possible failures and propose hardware implementation that minimizes catastrophic failures. Such hardware optimization bounds the possible logical value of the failed cells and allows us to compensate for the loss of accuracy via off-line training. By introducing the random weight defects during the training, we show that the model becomes more resilient on the device initialization failures, therefore, less prone to degrade the inference performance due to the failed devices. Our study sheds light on the hardware and software co-optimization procedure to cope with potentially catastrophic failures in the cross-bar array.<\/jats:p>","DOI":"10.1145\/3393669","type":"journal-article","created":{"date-parts":[[2020,7,7]],"date-time":"2020-07-07T12:38:34Z","timestamp":1594125514000},"page":"1-19","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Hardware and Software Co-optimization for the Initialization Failure of the ReRAM-based Cross-bar Array"],"prefix":"10.1145","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8486-9162","authenticated-orcid":false,"given":"Youngseok","family":"Kim","sequence":"first","affiliation":[{"name":"IBM Research, Albany, NY 12203, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Seyoung","family":"Kim","sequence":"additional","affiliation":[{"name":"Department of Materials Science and Engineering, POSTECH, Pohang, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chun-Chen","family":"Yeh","sequence":"additional","affiliation":[{"name":"IBM Research, Albany, NY 12203, USA and IBM Thomas J. Watson Research Center, Yorktown Heights, NY 10598, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Vijay","family":"Narayanan","sequence":"additional","affiliation":[{"name":"IBM Thomas J. Watson Research Center, Yorktown Heights, NY 10598, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jungwook","family":"Choi","sequence":"additional","affiliation":[{"name":"Department of Electronic Engineering, Hanyang University, Seoul, South Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2020,8,18]]},"reference":[{"key":"e_1_2_1_1_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-018-0180-5"},{"key":"e_1_2_1_2_1","doi-asserted-by":"crossref","unstructured":"G. Bersuker D. C. Gilmer and D. Veksler. 2019. Metal-oxide resistive random access memory (RRAM) technology: Material and operation details and ramifications. In Advances in Non-Volatile Memory and Storage Technology. Elsevier 35--102.  G. Bersuker D. C. Gilmer and D. Veksler. 2019. Metal-oxide resistive random access memory (RRAM) technology: Material and operation details and ramifications. In Advances in Non-Volatile Memory and Storage Technology. Elsevier 35--102.","DOI":"10.1016\/B978-0-08-102584-0.00002-4"},{"volume-title":"Proceedings of the IEEE International Electron Devices Meeting (IEDM\u201918)","year":"2018","author":"Bocquet M.","key":"e_1_2_1_3_1"},{"key":"e_1_2_1_4_1","doi-asserted-by":"publisher","DOI":"10.1109\/LED.2013.2288262"},{"key":"e_1_2_1_5_1","doi-asserted-by":"publisher","DOI":"10.1109\/TC.2014.12"},{"volume-title":"Prime: A novel processing-in-memory architecture for neural network computation in ReRAM-based main memory. In ACM SIGARCH Computer Architecture News","year":"2016","author":"Chi Ping","key":"e_1_2_1_6_1"},{"volume-title":"Vijayalakshmi Srinivasan, and Kailash Gopalakrishnan.","year":"2018","author":"Choi Jungwook","key":"e_1_2_1_7_1"},{"volume-title":"Binarized neural networks: Training deep neural networks with weights and activations constrained to +1 or \u22121. Arxiv Preprint Arxiv:1602.02830","year":"2016","author":"Courbariaux Matthieu","key":"e_1_2_1_8_1"},{"key":"e_1_2_1_9_1","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2017.00538"},{"volume-title":"Training LSTM networks with resistive cross-point devices. Arxiv Preprint Arxiv:1806.00166","year":"2018","author":"Gokmen Tayfun","key":"e_1_2_1_10_1"},{"key":"e_1_2_1_11_1","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2016.00333"},{"key":"e_1_2_1_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.sse.2015.10.003"},{"volume-title":"Proceedings of the IEEE International Electron Devices Meeting (IEDM\u201917)","year":"2017","author":"Guo X.","key":"e_1_2_1_13_1"},{"key":"e_1_2_1_14_1","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2018.2871057"},{"key":"e_1_2_1_15_1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_1_16_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46493-0_38"},{"key":"e_1_2_1_17_1","doi-asserted-by":"publisher","DOI":"10.1147\/JRD.2019.2947011"},{"key":"e_1_2_1_18_1","doi-asserted-by":"publisher","DOI":"10.1063\/1.1776636"},{"key":"e_1_2_1_19_1","doi-asserted-by":"publisher","DOI":"10.1063\/1.2959065"},{"key":"e_1_2_1_20_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41928-017-0002-z"},{"volume-title":"Proceedings of the 47th European Solid-State Device Research Conference (ESSDERC\u201917)","year":"2017","author":"Liu Rui","key":"e_1_2_1_21_1"},{"key":"e_1_2_1_22_1","first-page":"1","article-title":"ReRAM-based processing-in-memory architecture for recurrent neural network acceleration","volume":"99","author":"Long Yun","year":"2018","journal-title":"IEEE Trans. Very Large Scale Integr. (VLSI) Syst."},{"key":"e_1_2_1_23_1","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2017.2778940"},{"key":"e_1_2_1_24_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.mser.2014.06.002"},{"key":"e_1_2_1_25_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46493-0_32"},{"key":"e_1_2_1_26_1","doi-asserted-by":"publisher","DOI":"10.1039\/C8FD00107C"},{"key":"e_1_2_1_27_1","doi-asserted-by":"publisher","DOI":"10.1145\/3007787.3001139"},{"key":"e_1_2_1_28_1","doi-asserted-by":"publisher","DOI":"10.1109\/VTS.2011.5783775"},{"key":"e_1_2_1_29_1","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2017.55"},{"key":"e_1_2_1_30_1","first-page":"3","article-title":"XNOR-ReRAM: A scalable and parallel resistive synaptic architecture for binary neural networks","volume":"2","author":"Sun Xiaoyu","year":"2018","journal-title":"Algorithms"},{"volume-title":"The decline of computers as a general purpose technology: Why deep learning and the end of Moore\u2019s law are fragmenting computing. Retrieved from: SSRN 3287769","year":"2018","author":"Thompson Neil","key":"e_1_2_1_31_1"},{"key":"e_1_2_1_32_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2017.04.075"},{"key":"e_1_2_1_33_1","doi-asserted-by":"publisher","DOI":"10.1109\/JETCAS.2017.2776980"},{"key":"e_1_2_1_34_1","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2016-1446"},{"volume-title":"DoReFa-Net: Training low bitwidth convolutional neural networks with low bitwidth gradients. Arxiv Preprint Arxiv:1606.06160","year":"2016","author":"Zhou Shuchang","key":"e_1_2_1_35_1"},{"volume-title":"Proceedings of the IEEE International Electron Devices Meeting (IEDM\u201918)","year":"2018","author":"Zhou Z.","key":"e_1_2_1_36_1"},{"key":"e_1_2_1_37_1","unstructured":"Neta Zmora Guy Jacob and Gal Novik. 2018. Neural Network Distiller. DOI:https:\/\/doi.org\/10.5281\/zenodo.1297430  Neta Zmora Guy Jacob and Gal Novik. 2018. Neural Network Distiller. DOI:https:\/\/doi.org\/10.5281\/zenodo.1297430"}],"container-title":["ACM Journal on Emerging Technologies in Computing Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3393669","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3393669","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,17]],"date-time":"2025-06-17T22:41:17Z","timestamp":1750200077000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3393669"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,8,18]]},"references-count":37,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2020,10,31]]}},"alternative-id":["10.1145\/3393669"],"URL":"https:\/\/doi.org\/10.1145\/3393669","relation":{},"ISSN":["1550-4832","1550-4840"],"issn-type":[{"type":"print","value":"1550-4832"},{"type":"electronic","value":"1550-4840"}],"subject":[],"published":{"date-parts":[[2020,8,18]]},"assertion":[{"value":"2019-05-01","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2020-04-01","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2020-08-18","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}