{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T00:47:52Z","timestamp":1776386872985,"version":"3.51.2"},"reference-count":44,"publisher":"Proceedings of the National Academy of Sciences","issue":"41","license":[{"start":{"date-parts":[[2016,9,20]],"date-time":"2016-09-20T00:00:00Z","timestamp":1474329600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/www.pnas.org\/site\/misc\/userlicense.xhtml"}],"funder":[{"DOI":"10.13039\/100000185","name":"DOD | Defense Advanced Research Projects Agency","doi-asserted-by":"publisher","award":["FA9453-15-C-0055"],"award-info":[{"award-number":["FA9453-15-C-0055"]}],"id":[{"id":"10.13039\/100000185","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.pnas.org"],"crossmark-restriction":true},"short-container-title":["Proc. Natl. Acad. Sci. U.S.A."],"published-print":{"date-parts":[[2016,10,11]]},"abstract":"<jats:title>Significance<\/jats:title>\n          <jats:p>Brain-inspired computing seeks to develop new technologies that solve real-world problems while remaining grounded in the physical requirements of energy, speed, and size. Meeting these challenges requires high-performing algorithms that are capable of running on efficient hardware. Here, we adapt deep convolutional neural networks, which are today\u2019s state-of-the-art approach for machine perception in many domains, to perform classification tasks on neuromorphic hardware, which is today\u2019s most efficient platform for running neural networks. Using our approach, we demonstrate near state-of-the-art accuracy on eight datasets, while running at between 1,200 and 2,600 frames\/s and using between 25 and 275 mW.<\/jats:p>","DOI":"10.1073\/pnas.1604850113","type":"journal-article","created":{"date-parts":[[2016,9,21]],"date-time":"2016-09-21T02:40:49Z","timestamp":1474425649000},"page":"11441-11446","update-policy":"https:\/\/doi.org\/10.1073\/pnas.cm10313","source":"Crossref","is-referenced-by-count":546,"title":["Convolutional networks for fast, energy-efficient neuromorphic computing"],"prefix":"10.1073","volume":"113","author":[{"given":"Steven K.","family":"Esser","sequence":"first","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Paul A.","family":"Merolla","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"John V.","family":"Arthur","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Andrew S.","family":"Cassidy","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Rathinakumar","family":"Appuswamy","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Alexander","family":"Andreopoulos","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"David J.","family":"Berg","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Jeffrey L.","family":"McKinstry","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Timothy","family":"Melano","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Davis R.","family":"Barch","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Carmelo","family":"di Nolfo","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Pallab","family":"Datta","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Arnon","family":"Amir","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Brian","family":"Taba","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Myron D.","family":"Flickner","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]},{"given":"Dharmendra S.","family":"Modha","sequence":"additional","affiliation":[{"name":"Brain-Inspired Computing, IBM Research\u2013Almaden, San Jose, CA 95120"}]}],"member":"341","published-online":{"date-parts":[[2016,9,20]]},"reference":[{"key":"e_1_3_3_1_2","doi-asserted-by":"publisher","DOI":"10.1038\/323533a0"},{"key":"e_1_3_3_2_2","doi-asserted-by":"publisher","DOI":"10.1007\/BF00344251"},{"key":"e_1_3_3_3_2","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1989.1.4.541"},{"key":"e_1_3_3_4_2","doi-asserted-by":"publisher","DOI":"10.1038\/nature14539"},{"key":"e_1_3_3_5_2","doi-asserted-by":"publisher","DOI":"10.1109\/5.58356"},{"key":"e_1_3_3_6_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.1254642"},{"key":"e_1_3_3_7_2","unstructured":"M Lin Q Chen S Yan Network in network. International Conference on Learning Representations (ICLR) . arXiv:1312.4400v3. (2014)."},{"key":"e_1_3_3_8_2","doi-asserted-by":"publisher","DOI":"10.7554\/eLife.10778"},{"key":"e_1_3_3_9_2","unstructured":"J Jin A Dundar E Culurciello Flattened convolutional neural networks for feedforward acceleration. arXiv : 1412.5474. (2014)."},{"key":"e_1_3_3_10_2","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2015.00222"},{"key":"e_1_3_3_11_2","first-page":"3105","article-title":"Binaryconnect: Training deep neural networks with binary weights during propagations","volume":"28","author":"Courbariaux M","year":"2015","unstructured":"M Courbariaux, Y Bengio, JP David, Binaryconnect: Training deep neural networks with binary weights during propagations. Advances in Neural Information Processing Systems 28, 3105\u20133113 (2015).","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_3_12_2","unstructured":"Z Wu D Lin X Tang Adjustable bounded rectifiers: Towards deep binary representations. arXiv:1511.06201. (2015)."},{"key":"e_1_3_3_13_2","doi-asserted-by":"crossref","unstructured":"PU Diehl Truehappiness: Neuromorphic emotion recognition on truenorth. arXiv:1601.04183. (2016).","DOI":"10.1109\/IJCNN.2016.7727758"},{"key":"e_1_3_3_14_2","first-page":"1135","article-title":"Learning both weights and connections for efficient neural network","volume":"28","author":"Han S","year":"2015","unstructured":"S Han, J Pool, J Tran, W Dally, Learning both weights and connections for efficient neural network. Advances in Neural Information Processing Systems 28, 1135\u20131143 (2015).","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_3_15_2","first-page":"1117","article-title":"Backpropagation for energy-efficient neuromorphic computing","volume":"28","author":"Esser SK","year":"2015","unstructured":"SK Esser, R Appuswamy, P Merolla, JV Arthur, DS Modha, Backpropagation for energy-efficient neuromorphic computing. Advances in Neural Information Processing Systems 28, 1117\u20131125 (2015).","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_3_16_2","unstructured":"M Courbariaux Y Bengio Binarynet: Training deep neural networks with weights and activations constrained to+ 1 or-1. arXiv:1602.02830. (2016)."},{"key":"e_1_3_3_17_2","doi-asserted-by":"publisher","DOI":"10.1016\/S0042-6989(97)00121-1"},{"key":"e_1_3_3_18_2","volume-title":"Proceeding of the 23rd ACM International Conference on Multimedia","author":"Vedaldi A","year":"2015","unstructured":"A Vedaldi, K Lenc, MatConvNet: Convolutional neural networks for MATLAB. Proceeding of the 23rd ACM International Conference on Multimedia (ACM, New York, 2015)."},{"key":"e_1_3_3_19_2","first-page":"1","volume-title":"Neural Networks (IJCNN), The 2013 International Joint Conference","author":"Amir A","year":"2013","unstructured":"A Amir, , Cognitive computing programming paradigm: A corelet language for composing networks of neurosynaptic cores. Neural Networks (IJCNN), The 2013 International Joint Conference (IEEE, Dallas), pp. 1\u201310 (2013)."},{"key":"e_1_3_3_20_2","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2013.2259038"},{"key":"e_1_3_3_21_2","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2013.00011"},{"key":"e_1_3_3_22_2","doi-asserted-by":"publisher","DOI":"10.1109\/TBCAS.2013.2255873"},{"key":"e_1_3_3_23_2","doi-asserted-by":"crossref","first-page":"675","DOI":"10.1109\/BioCAS.2014.6981816","volume-title":"Proceedings of the Biomedical Circuits and Systems Conference (BioCAS), 2014 IEEE","author":"Park J","year":"2014","unstructured":"J Park, S Ha, T Yu, E Neftci, G Cauwenberghs, A 65k-neuron 73-Mevents\/s 22-pJ\/event asynchronous micro-pipelined integrate-and-fire array transceiver. Proceedings of the Biomedical Circuits and Systems Conference (BioCAS), 2014 IEEE (IEEE, Lausanne, Switzerland), pp. 675\u2013678 (2014)."},{"key":"e_1_3_3_24_2","unstructured":"S Ioffe C Szegedy Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv:1502.03167. (2015)."},{"key":"e_1_3_3_25_2","first-page":"1097","article-title":"Imagenet classification with deep convolutional neural networks","volume":"25","author":"Krizhevsky A","year":"2012","unstructured":"A Krizhevsky, I Sutskever, GE Hinton, Imagenet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems 25, 1097\u20131105 (2012).","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_3_26_2","volume-title":"Learning Multiple Layers of Features From Tiny Images","author":"Krizhevsky A","year":"2009","unstructured":"A Krizhevsky Learning Multiple Layers of Features From Tiny Images (University of Toronto, Toronto, 2009)."},{"key":"e_1_3_3_27_2","volume-title":"NIPS Workshop on Deep Learning and Unsupervised Feature Learning 2011","author":"Netzer Y","year":"2011","unstructured":"Y Netzer, , Reading digits in natural images with unsupervised feature learning. NIPS Workshop on Deep Learning and Unsupervised Feature Learning 2011 (NIPS, Granada, Spain, 2011)."},{"key":"e_1_3_3_28_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2012.02.016"},{"key":"e_1_3_3_29_2","first-page":"25","volume-title":"Proceedings of the 1st ACM International Conference on Multimedia Retrieval","author":"Romberg S","year":"2011","unstructured":"S Romberg, LG Pueyo, R Lienhart, RV Zwol, Scalable logo recognition in real-world images. Proceedings of the 1st ACM International Conference on Multimedia Retrieval (ACM, New York), pp. 25 (2011)."},{"key":"e_1_3_3_30_2","volume-title":"TIMIT Acoustic-Phonetic Continuous Speech Corpus LDC93S1","author":"Garofolo J","year":"1993","unstructured":"J Garofolo, TIMIT Acoustic-Phonetic Continuous Speech Corpus LDC93S1 (Linguistic Data Consortium, Philadelphia, 1993)."},{"key":"e_1_3_3_31_2","doi-asserted-by":"publisher","DOI":"10.1016\/0167-6393(93)90095-3"},{"key":"e_1_3_3_32_2","unstructured":"JT Springenberg A Dosovitskiy T Brox M Riedmiller Striving for simplicity: The all convolutional net. arXiv:1412.6806. (2014)."},{"key":"e_1_3_3_33_2","first-page":"54","volume-title":"Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis","author":"Preissl R","year":"2012","unstructured":"R Preissl, , Compass: A scalable simulator for an architecture for cognitive computing. Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis (IEEE Computer Society Press, Los Alamitos, CA), pp. 54 (2012)."},{"key":"e_1_3_3_34_2","unstructured":"CY Lee S Xie P Gallagher Z Zhang Z Tu Deeply-supervised nets. arXiv:1409.5185. (2014)."},{"key":"e_1_3_3_35_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2012.02.023"},{"key":"e_1_3_3_36_2","first-page":"1","article-title":"Using artificial neural network for robust voice activity detection under adverse conditions","volume":"1","author":"Pham TV","year":"2009","unstructured":"TV Pham, CT Tang, M Stadtschnitzer, Using artificial neural network for robust voice activity detection under adverse conditions. Proceedings of the International Conference on Computing and Communication Technologies 1, 1\u20138 (2009).","journal-title":"Proceedings of the International Conference on Computing and Communication Technologies"},{"key":"e_1_3_3_37_2","volume-title":"Proceedings of the IEEE Workshop on Automatic Speech Recognition and Understanding","author":"Chang HA","year":"2007","unstructured":"HA Chang, JR Glass, Hierarchical large-margin gaussian mixture models for phonetic classification. Proceedings of the IEEE Workshop on Automatic Speech Recognition and Understanding (IEEE, Kyoto, 2007)."},{"key":"e_1_3_3_38_2","first-page":"273","volume-title":"Proceedings of the IEEE Workshop on Automatic Speech Recognition and Understanding","author":"Graves A","year":"2013","unstructured":"A Graves, N Jaitly, A Mohamed, Hybrid speech recognition with deep bidirectional lstm. Proceedings of the IEEE Workshop on Automatic Speech Recognition and Understanding (IEEE, Olomouc, Czech Republic), pp. 273\u2013278 (2013)."},{"key":"e_1_3_3_39_2","first-page":"1","volume-title":"Proceedings of the IEEE Custom Integrated Circuits Conference","author":"Merolla P","year":"2011","unstructured":"P Merolla, , A digital neurosynaptic core using embedded crossbar memory with 45pJ per spike in 45nm. Proceedings of the IEEE Custom Integrated Circuits Conference (IEEE, San Jose, CA), pp. 1\u20134 (2011)."},{"key":"e_1_3_3_40_2","first-page":"1","volume-title":"Proceedings of the Neural Networks (IJCNN), The 2013 International Joint Conference","author":"Esser SK","year":"2013","unstructured":"SK Esser, , Cognitive computing systems: Algorithms and applications for networks of neurosynaptic cores. Proceedings of the Neural Networks (IJCNN), The 2013 International Joint Conference (IEEE, Dallas), pp. 1\u201310 (2013)."},{"key":"e_1_3_3_41_2","doi-asserted-by":"crossref","unstructured":"PU Diehl G Zarrella A Cassidy BU Pedroni E Neftci Conversion of artificial recurrent neural networks to spiking neural networks for low-power neuromorphic hardware. arXiv:1601.04187. (2016).","DOI":"10.1109\/ICRC.2016.7738691"},{"key":"e_1_3_3_42_2","volume-title":"Proceedings of the IEEE International Symposium on Circuits and Systems","author":"Das S","year":"2015","unstructured":"S Das, , Gibbs sampling with low-power spiking digital neurons. Proceedings of the IEEE International Symposium on Circuits and Systems (IEEE, Lisbon, Portugal, 2015)."},{"key":"e_1_3_3_43_2","first-page":"683","volume-title":"Proceedings of the 2015 Design, Automation & Test in Europe Conference & Exhibition","author":"Conti F","year":"2015","unstructured":"F Conti, L Benini, A ultra-low-energy convolution engine for fast brain-inspired vision in multicore clusters. Proceedings of the 2015 Design, Automation & Test in Europe Conference & Exhibition (IEEE, Grenoble, France), pp. 683\u2013688 (2015)."},{"key":"e_1_3_3_44_2","doi-asserted-by":"crossref","first-page":"262","DOI":"10.1109\/ISSCC.2016.7418007","volume-title":"Proceedings of the 2016 IEEE International Solid-State Circuits Conference (ISSCC)","author":"Chen YH","year":"2016","unstructured":"YH Chen, T Krishna, J Emer, V Sze, 14.5 Eyeriss: An energy-efficient reconfigurable accelerator for deep convolutional neural networks. Proceedings of the 2016 IEEE International Solid-State Circuits Conference (ISSCC) (IEEE, San Francisco), pp. 262\u2013263 (2016)."}],"container-title":["Proceedings of the National Academy of Sciences"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/www.pnas.org\/syndication\/doi\/10.1073\/pnas.1604850113","content-type":"unspecified","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pnas.org\/doi\/pdf\/10.1073\/pnas.1604850113","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,4,13]],"date-time":"2022-04-13T05:28:01Z","timestamp":1649827681000},"score":1,"resource":{"primary":{"URL":"https:\/\/pnas.org\/doi\/full\/10.1073\/pnas.1604850113"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2016,9,20]]},"references-count":44,"journal-issue":{"issue":"41","published-print":{"date-parts":[[2016,10,11]]}},"alternative-id":["10.1073\/pnas.1604850113"],"URL":"https:\/\/doi.org\/10.1073\/pnas.1604850113","relation":{},"ISSN":["0027-8424","1091-6490"],"issn-type":[{"value":"0027-8424","type":"print"},{"value":"1091-6490","type":"electronic"}],"subject":[],"published":{"date-parts":[[2016,9,20]]},"assertion":[{"value":"2016-09-20","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}