{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,13]],"date-time":"2026-06-13T17:33:05Z","timestamp":1781371985019,"version":"3.54.1"},"reference-count":42,"publisher":"Wiley","issue":"8","license":[{"start":{"date-parts":[[2021,5,20]],"date-time":"2021-05-20T00:00:00Z","timestamp":1621468800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100000781","name":"European Research Council","doi-asserted-by":"publisher","award":["715872"],"award-info":[{"award-number":["715872"]}],"id":[{"id":"10.13039\/501100000781","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["advanced.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Advanced Intelligent Systems"],"published-print":{"date-parts":[[2021,8]]},"abstract":"<jats:sec><jats:label\/><jats:p>Neural networks cannot typically be trained locally in edge\u2010computing systems due to severe energy constraints. It has, therefore, become commonplace to train them \u201cex situ\u201d and transfer the resulting model to a dedicated inference hardware. Resistive memory arrays are of particular interest for realizing such inference hardware, because they offer an extremely low\u2010power implementation of the dot\u2010product operation. However, the transfer of high\u2010precision software parameters to the imprecise and random conductance states of resistive memories poses significant challenges. Here, it is proposed that Bayesian neural networks can be more suitable for model transfer, because, such as device conductance states, their parameters are described by random variables. The ex situ training of a Bayesian neural network is performed, and then, the resulting software model is transferred in a single programming step to an array of 16\u2009384 resistive memory devices. On an illustrative classification task, it is observed that the transferred decision boundaries and the prediction uncertainties of the software model are well preserved. This work demonstrates that resistive memory\u2010based Bayesian neural networks are a promising direction in the development of resistive memory compatible edge inference hardware.<\/jats:p><\/jats:sec>","DOI":"10.1002\/aisy.202000103","type":"journal-article","created":{"date-parts":[[2021,5,20]],"date-time":"2021-05-20T11:09:54Z","timestamp":1621508994000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":29,"title":["Ex Situ Transfer of Bayesian Neural Networks to Resistive Memory\u2010Based Inference Hardware"],"prefix":"10.1002","volume":"3","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0326-2121","authenticated-orcid":false,"given":"Thomas","family":"Dalgaty","sequence":"first","affiliation":[{"name":"CEA-LETI Universit\u00e9 Grenoble Alpes  Grenoble 38000 France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eduardo","family":"Esmanhotto","sequence":"additional","affiliation":[{"name":"CEA-LETI Universit\u00e9 Grenoble Alpes  Grenoble 38000 France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Niccolo","family":"Castellani","sequence":"additional","affiliation":[{"name":"CEA-LETI Universit\u00e9 Grenoble Alpes  Grenoble 38000 France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Damien","family":"Querlioz","sequence":"additional","affiliation":[{"name":"Universit\u00e9 Paris-Saclay CNRS Centre de Nanosciences et de Nanotechnologies  Palaiseau 91120 France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Elisa","family":"Vianello","sequence":"additional","affiliation":[{"name":"CEA-LETI Universit\u00e9 Grenoble Alpes  Grenoble 38000 France"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2021,5,20]]},"reference":[{"key":"e_1_2_5_2_1","doi-asserted-by":"publisher","DOI":"10.1038\/nature14539"},{"key":"e_1_2_5_3_1","first-page":"1097","volume-title":"Advances in Neural Information Processing Systems 25","author":"Krizhevsky A.","year":"2012"},{"key":"e_1_2_5_4_1","doi-asserted-by":"publisher","DOI":"10.1038\/nature14236"},{"key":"e_1_2_5_5_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-019-1724-z"},{"key":"e_1_2_5_6_1","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2016.2579198"},{"key":"e_1_2_5_7_1","unstructured":"E.Strubell A.Ganesh A.Mccallum in57th Annual Meeting of the Association for Computational Linguistics (ACL) Florence Italy July 2019."},{"key":"e_1_2_5_8_1","unstructured":"D.Li X.Chen M.Becchi Z.Zong in2016 IEEE International Conferences on Big Data and Cloud Computing (BDCloud) Social Computing and Networking (SocialCom) Sustainable Computing and Communications (SustainCom) (BDCloud-SocialCom-SustainCom) IEEE Piscataway NJ2016 pp.477\u2013484."},{"key":"e_1_2_5_9_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41928-018-0059-3"},{"key":"e_1_2_5_10_1","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2918951"},{"key":"e_1_2_5_11_1","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2010.2070050"},{"key":"e_1_2_5_12_1","doi-asserted-by":"publisher","DOI":"10.1038\/nmat2024"},{"key":"e_1_2_5_13_1","doi-asserted-by":"publisher","DOI":"10.1002\/adma.201290080"},{"key":"e_1_2_5_14_1","doi-asserted-by":"publisher","DOI":"10.1063\/1.126902"},{"key":"e_1_2_5_15_1","doi-asserted-by":"crossref","unstructured":"M.Hu J. P.Strachan Z.Li R.Stanley Williams in2016 17th Int. Symp. on Quality Electronic Design (ISQED) Santa Clara CA USA2016 pp.374\u2013379.","DOI":"10.1109\/ISQED.2016.7479230"},{"key":"e_1_2_5_16_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41563-019-0291-x"},{"key":"e_1_2_5_17_1","doi-asserted-by":"crossref","unstructured":"R.Mochida K.Kouno Y.Hayata M.Nakayama T.Ono H.Suwa R.Yasuhara K.Katayama T.Mikawa Y.Gohou in2018 IEEE Symp. on VLSI Technology Honolulu HI USA2018 pp.175\u2013176.","DOI":"10.1109\/VLSIT.2018.8510676"},{"key":"e_1_2_5_18_1","doi-asserted-by":"publisher","DOI":"10.1002\/aelm.201900026"},{"key":"e_1_2_5_19_1","doi-asserted-by":"publisher","DOI":"10.1002\/adma.201705914"},{"key":"e_1_2_5_20_1","unstructured":"A.Valentian F.Rummens E.Vianello T.Mesquida C.Lacat-Mathieu de Boissac C.Reita inInternational Electron Device Meeting SanFrancisco CA USA2019."},{"key":"e_1_2_5_21_1","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2015.2437616"},{"key":"e_1_2_5_22_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-018-04933-y"},{"key":"e_1_2_5_23_1","doi-asserted-by":"crossref","unstructured":"M.Bocquet T.Hirztlin J.-O.Klein E.Nowak E.Vianello J.-M.Portal D.Querlioz in2018 IEEE International Electron Devices Meeting (IEDM) IEEE Piscataway NJ2018 pp.20\u201326.","DOI":"10.1109\/IEDM.2018.8614639"},{"key":"e_1_2_5_24_1","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2019.01383"},{"key":"e_1_2_5_25_1","doi-asserted-by":"publisher","DOI":"10.1063\/1.5108663"},{"key":"e_1_2_5_26_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41928-020-00523-3"},{"key":"e_1_2_5_27_1","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2016.2546064"},{"key":"e_1_2_5_28_1","doi-asserted-by":"publisher","DOI":"10.1063\/1.5021332"},{"key":"e_1_2_5_29_1","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2016.00104"},{"key":"e_1_2_5_30_1","doi-asserted-by":"crossref","unstructured":"P.Mroszczyk P.Dudek in2014 IEEE International Symposium on Circuits and Systems (ISCAS) IEEE Piscataway NJ2014 pp.1576\u20131579.","DOI":"10.1109\/ISCAS.2014.6865450"},{"key":"e_1_2_5_31_1","doi-asserted-by":"publisher","DOI":"10.1038\/nature14541"},{"key":"e_1_2_5_32_1","volume-title":"Bayesian Learning for Neural Networks","author":"Neal R. M.","year":"2012"},{"key":"e_1_2_5_33_1","doi-asserted-by":"crossref","unstructured":"A.Grossi E.Nowak C.Zambelli C.Pellissier S.Bernasconi G.Cibrario K. E.Hajjam R.Crochemore J. F.Nodin P.Olivo L.Perniola in2016 IEEE Int. Electron Devices Meeting (IEDM) IEEE Piscataway NJ2016 pp.4.7.1\u20134.7.4.","DOI":"10.1109\/IEDM.2016.7838348"},{"key":"e_1_2_5_34_1","doi-asserted-by":"publisher","DOI":"10.1109\/TED.2011.2167513"},{"key":"e_1_2_5_35_1","doi-asserted-by":"publisher","DOI":"10.1109\/79.543975"},{"key":"e_1_2_5_36_1","doi-asserted-by":"publisher","DOI":"10.1109\/TED.2019.2894387"},{"key":"e_1_2_5_37_1","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.33.2.25"},{"key":"e_1_2_5_38_1","first-page":"430","volume":"18","author":"Kucukelbir A.","year":"2017","journal-title":"J. Mach. Learn. Res."},{"key":"e_1_2_5_39_1","first-page":"1351","volume":"15","author":"Hoffman M.","year":"2014","journal-title":"J. Mach. Learn. Res."},{"key":"e_1_2_5_40_1","volume-title":"Bayesian Deep Learning","author":"Wiecki T.","year":"2016"},{"key":"e_1_2_5_41_1","doi-asserted-by":"crossref","unstructured":"D. D.Fan J.Nguyen R.Thakker N.Alatur A.-A.Agha-Mohammadi E. A.Theodorou 2020 IEEE International Conference on Robotics and Automation (ICRA) Paris France2019 pp.4093\u20134099.","DOI":"10.1109\/ICRA40945.2020.9196709"},{"key":"e_1_2_5_42_1","doi-asserted-by":"publisher","DOI":"10.1056\/NEJM200002103420601"},{"key":"e_1_2_5_43_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jacc.2007.09.073"}],"container-title":["Advanced Intelligent Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/aisy.202000103","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/aisy.202000103","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/advanced.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/aisy.202000103","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,7]],"date-time":"2025-10-07T19:02:36Z","timestamp":1759863756000},"score":1,"resource":{"primary":{"URL":"https:\/\/advanced.onlinelibrary.wiley.com\/doi\/10.1002\/aisy.202000103"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,5,20]]},"references-count":42,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2021,8]]}},"alternative-id":["10.1002\/aisy.202000103"],"URL":"https:\/\/doi.org\/10.1002\/aisy.202000103","archive":["Portico"],"relation":{},"ISSN":["2640-4567","2640-4567"],"issn-type":[{"value":"2640-4567","type":"print"},{"value":"2640-4567","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,5,20]]},"assertion":[{"value":"2020-05-15","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-05-20","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"2000103"}}