{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,17]],"date-time":"2026-07-17T19:05:39Z","timestamp":1784315139379,"version":"3.55.0"},"reference-count":63,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"4","license":[{"start":{"date-parts":[[2021,4,1]],"date-time":"2021-04-01T00:00:00Z","timestamp":1617235200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2021,4,1]],"date-time":"2021-04-01T00:00:00Z","timestamp":1617235200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2021,4,1]],"date-time":"2021-04-01T00:00:00Z","timestamp":1617235200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["61771201"],"award-info":[{"award-number":["61771201"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Guangdong Introducing Innovative and Enterpreneurial Teams","award":["2017ZT07X183"],"award-info":[{"award-number":["2017ZT07X183"]}]},{"DOI":"10.13039\/501100000923","name":"Australian Research Council","doi-asserted-by":"publisher","award":["DP180103424"],"award-info":[{"award-number":["DP180103424"]}],"id":[{"id":"10.13039\/501100000923","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000923","name":"Australian Research Council","doi-asserted-by":"publisher","award":["DE190101473"],"award-info":[{"award-number":["DE190101473"]}],"id":[{"id":"10.13039\/501100000923","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000923","name":"Australian Research Council","doi-asserted-by":"publisher","award":["FL-170100117"],"award-info":[{"award-number":["FL-170100117"]}],"id":[{"id":"10.13039\/501100000923","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Pattern Anal. Mach. Intell."],"published-print":{"date-parts":[[2021,4,1]]},"DOI":"10.1109\/tpami.2019.2948352","type":"journal-article","created":{"date-parts":[[2019,10,21]],"date-time":"2019-10-21T23:40:45Z","timestamp":1571701245000},"page":"1352-1368","source":"Crossref","is-referenced-by-count":70,"title":["Orthogonal Deep Neural Networks"],"prefix":"10.1109","volume":"43","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-4298-5165","authenticated-orcid":false,"given":"Shuai","family":"Li","sequence":"first","affiliation":[{"name":"School of Electronic and Information Engineering, South China University of Technology, Guangzhou, Guangdong, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2661-5700","authenticated-orcid":false,"given":"Kui","family":"Jia","sequence":"additional","affiliation":[{"name":"School of Electronic and Information Engineering, South China University of Technology, Guangzhou, Guangdong, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3719-9001","authenticated-orcid":false,"given":"Yuxin","family":"Wen","sequence":"additional","affiliation":[{"name":"School of Electronic and Information Engineering, South China University of Technology, Guangzhou, Guangdong, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9640-6472","authenticated-orcid":false,"given":"Tongliang","family":"Liu","sequence":"additional","affiliation":[{"name":"Faculty of Engineering, UBTECH Sydney AI Centre and the School of Computer Science, University of Sydney, Darlington, NSW, Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7225-5449","authenticated-orcid":false,"given":"Dacheng","family":"Tao","sequence":"additional","affiliation":[{"name":"Faculty of Engineering, UBTECH Sydney AI Centre and the School of Computer Science, University of Sydney, Darlington, NSW, Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref39","first-page":"562","article-title":"Deeply-supervised nets","author":"lee","year":"2015","journal-title":"Proc 18th Int Conf Artif Intell Statist"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1007\/3-540-49430-8_2"},{"key":"ref33","article-title":"Adam: A method for stochastic optimization","author":"kingma","year":"2015","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref32","article-title":"On large-batch training for deep learning: Generalization gap and sharp minima","author":"keskar","year":"2017","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref31","article-title":"Generalization in deep learning","author":"kawaguchi","year":"2017","journal-title":"CoRR"},{"key":"ref30","first-page":"586","article-title":"Deep learning without poor local minima","author":"kawaguchi","year":"2016","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1109\/5.726791"},{"key":"ref36","first-page":"2815","article-title":"Data-dependent stability of stochastic gradient descent","author":"kuzborskij","year":"0","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref35","article-title":"Learning multiple layers of features from tiny images","author":"krizhevsky","year":"2009"},{"key":"ref34","first-page":"227","article-title":"$\\epsilon$?-entropy and $\\epsilon$?-capacity of sets in functional spaces","volume":"17","author":"kolmogorov","year":"2002","journal-title":"Amer Math Soc Translations"},{"key":"ref60","article-title":"Global optimality conditions for deep neural networks","author":"yun","year":"2018","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref62","article-title":"Understanding deep learning requires rethinking generalization","author":"zhang","year":"2017","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.5244\/C.30.87"},{"key":"ref63","article-title":"Theory of deep learning IIB: Optimization properties of SGD","author":"zhang","year":"2018","journal-title":"CoRR"},{"key":"ref28","first-page":"448","article-title":"Batch normalization: Accelerating deep network training by reducing internal covariate shift","author":"ioffe","year":"2015","journal-title":"Proc 32nd Int Conf Mach Learn"},{"key":"ref27","first-page":"1333","article-title":"Discriminative robust transformation learning","author":"huang","year":"2015","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.425"},{"key":"ref2","first-page":"1120","article-title":"Unitary evolution recurrent neural networks","author":"arjovsky","year":"0","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref1","author":"absil","year":"2007","journal-title":"Optimization Algorithms on Matrix Manifolds"},{"key":"ref20","first-page":"1225","article-title":"Train faster, generalize better: Stability of stochastic gradient descent","volume":"48","author":"hardt","year":"2016","journal-title":"Proc 33rd Int Conf Mach Learn"},{"key":"ref22","first-page":"630","article-title":"Identity mappings in deep residual networks","author":"he","year":"2016","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref24","article-title":"Improving neural networks by preventing co-adaptation of feature detectors","author":"hinton","year":"2012","journal-title":"CoRR"},{"key":"ref23","article-title":"Benchmarking neural network robustness to common corruptions and perturbations","author":"hendrycks","year":"2019","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.243"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1016\/0893-6080(89)90020-8"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/TSP.2017.2708039"},{"key":"ref51","first-page":"1139","article-title":"On the importance of initialization and momentum in deep learning","volume":"28","author":"sutskever","year":"2013","journal-title":"Proc 30th Int Conf Mach Learn"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1007\/s10994-011-5268-1"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.634"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.539"},{"key":"ref56","first-page":"4887","article-title":"Full-capacity unitary recurrent neural networks","author":"wisdom","year":"0","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref55","first-page":"718","article-title":"Analysis of deep neural networks with extended data jacobian matrix","author":"wang","year":"2016","journal-title":"Proc 33nd Int Conf Mach Learn"},{"key":"ref54","first-page":"351","article-title":"Dropout training as adaptive regularization","author":"wager","year":"2013","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref53","first-page":"2415","article-title":"Distance Preserving Embeddings for General n-Dimensional Manifolds","volume":"14","author":"verma","year":"2013","journal-title":"J Mach Learn Res"},{"key":"ref52","article-title":"Examining the impact of blur on recognition by convolutional networks","author":"vasiljevic","year":"2016","journal-title":"arXiv 1611 05760"},{"key":"ref10","first-page":"499","article-title":"Stability and generalization","volume":"2","author":"bousquet","year":"2002","journal-title":"J Mach Learn Res"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.2005.858979"},{"key":"ref40","first-page":"2159","article-title":"Algorithmic stability and hypothesis complexity","author":"liu","year":"2017","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref12","first-page":"854","article-title":"Parseval networks: Improving robustness to adversarial examples","author":"cisse","year":"2017","journal-title":"Proc 34th Int Conf Mach Learn"},{"key":"ref13","first-page":"2933","article-title":"Identifying and attacking the saddle point problem in high-dimensional non-convex optimization","author":"dauphin","year":"2014","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref14","first-page":"1019","article-title":"Sharp minima can generalize for deep nets","author":"dinh","year":"2017","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/ICCCN.2017.8038465"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1007\/BF00344251"},{"key":"ref17","first-page":"7548","article-title":"Generalisation in humans and deep neural networks","author":"geirhos","year":"2018","journal-title":"Advances Neural Inf Process Syst"},{"key":"ref18","first-page":"249","article-title":"Understanding the difficulty of training deep feedforward neural networks","author":"glorot","year":"2010","journal-title":"Proc Int Conf Artif Intell Statist"},{"key":"ref19","first-page":"315","article-title":"Deep sparse rectifier neural networks","author":"glorot","year":"2011","journal-title":"Proc Int Conf Artif Intell Statist"},{"key":"ref4","first-page":"4261","article-title":"Can we gain more from orthogonality regularizations in training deep networks?","author":"bansal","year":"2018","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref3","first-page":"2814","article-title":"Understanding dropout","author":"baldi","year":"2013","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/18.256500"},{"key":"ref5","first-page":"4266","article-title":"Can we gain more from orthogonality regularizations in training deep CNNs?","author":"bansal","year":"2018","journal-title":"Proc 32nd Int Conf Neural Inf Process Syst"},{"key":"ref8","article-title":"On the effect of pooling on the geometry of representations","author":"b\u00e9cigneul","year":"2017"},{"key":"ref7","first-page":"6241","article-title":"Spectrally-normalized margin bounds for neural networks","author":"bartlett","year":"2017","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref49","article-title":"Very deep convolutional networks for large-scale image recognition","author":"simonyan","year":"2015","journal-title":"Proc Int Conf Learn Representation"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/TAC.2013.2254619"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0816-y"},{"key":"ref45","first-page":"4788","article-title":"Resurrecting the sigmoid in deep learning through dynamical isometry: Theory and practice","author":"pennington","year":"2017","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref48","article-title":"The singular values of convolutional layers","author":"sedghi","year":"2019","journal-title":"Proc Int Conf Learn Representation"},{"key":"ref47","article-title":"Exact solutions to the nonlinear dynamics of learning in deep linear neural networks","author":"saxe","year":"2014","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref42","first-page":"2924","article-title":"On the number of linear regions of deep neural networks","author":"mont\u00fafar","year":"2014","journal-title":"Proc Advances Neural Inf Process Syst"},{"key":"ref41","author":"mohri","year":"2012","journal-title":"Foundations of Machine Learning"},{"key":"ref44","article-title":"Optimization on submanifolds of convolution kernels in CNNs","author":"ozay","year":"2016","journal-title":"CoRR"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-662-02772-1"}],"container-title":["IEEE Transactions on Pattern Analysis and Machine Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/34\/9370031\/08877742.pdf?arnumber=8877742","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,1,15]],"date-time":"2024-01-15T21:10:30Z","timestamp":1705353030000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/8877742\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,4,1]]},"references-count":63,"journal-issue":{"issue":"4"},"URL":"https:\/\/doi.org\/10.1109\/tpami.2019.2948352","relation":{},"ISSN":["0162-8828","2160-9292","1939-3539"],"issn-type":[{"value":"0162-8828","type":"print"},{"value":"2160-9292","type":"electronic"},{"value":"1939-3539","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,4,1]]}}}