{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,2]],"date-time":"2026-05-02T15:21:01Z","timestamp":1777735261602,"version":"3.51.4"},"reference-count":36,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2024,1,1]],"date-time":"2024-01-01T00:00:00Z","timestamp":1704067200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2024,1,1]],"date-time":"2024-01-01T00:00:00Z","timestamp":1704067200000},"content-version":"am","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2024,1,1]],"date-time":"2024-01-01T00:00:00Z","timestamp":1704067200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2024,1,1]],"date-time":"2024-01-01T00:00:00Z","timestamp":1704067200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"NSF","award":["CCF-2240708"],"award-info":[{"award-number":["CCF-2240708"]}]},{"name":"NSF","award":["CCF-2241298"],"award-info":[{"award-number":["CCF-2241298"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Signal Process. Lett."],"published-print":{"date-parts":[[2024]]},"DOI":"10.1109\/lsp.2024.3374085","type":"journal-article","created":{"date-parts":[[2024,3,6]],"date-time":"2024-03-06T19:08:11Z","timestamp":1709752091000},"page":"795-799","source":"Crossref","is-referenced-by-count":1,"title":["Convergence Analysis for Learning Orthonormal Deep Linear Neural Networks"],"prefix":"10.1109","volume":"31","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-0388-8840","authenticated-orcid":false,"given":"Zhen","family":"Qin","sequence":"first","affiliation":[{"name":"Department of Computer Science and Engineering, Ohio State University, Columbus, OH, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2948-9535","authenticated-orcid":false,"given":"Xuwei","family":"Tan","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, Ohio State University, Columbus, OH, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3856-0375","authenticated-orcid":false,"given":"Zhihui","family":"Zhu","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, Ohio State University, Columbus, OH, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","article-title":"All you need is a good init","author":"Mishkin","year":"2015"},{"key":"ref2","article-title":"Exact solutions to the nonlinear dynamics of learning in deep linear neural networks","author":"Saxe","year":"2013"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/72.279181"},{"key":"ref4","first-page":"1310","article-title":"On the difficulty of training recurrent neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Pascanu","year":"2013"},{"key":"ref5","article-title":"A simple way to initialize recurrent networks of rectified linear units","author":"Le","year":"2015"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.5555\/3045390.3045509"},{"key":"ref7","first-page":"580","article-title":"Which neural net architectures give rise to exploding and vanishing gradients","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"31","author":"Hanin","year":"2018"},{"key":"ref8","article-title":"Generalized backpropagation Etude De Cas: Orthogonality","author":"Harandi","year":"2016"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2019.2948352"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00646"},{"key":"ref11","first-page":"1","article-title":"Efficient Riemannian optimization on the Stiefel manifold via the cayley transform","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Li","year":"2020"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.01152"},{"key":"ref13","first-page":"15743","article-title":"Existence, stability and scalability of orthogonal convolutional neural networks","volume":"23","author":"Malgouyres","year":"2022","journal-title":"J. Mach. Learn. Res."},{"key":"ref14","article-title":"DizzyRNN: Reparameterizing recurrent neural networks for norm-preserving backpropagation","author":"Dorobantu","year":"2016"},{"key":"ref15","first-page":"2401","article-title":"Efficient orthogonal parametrisation of recurrent neural networks using householder reflections","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Mhammedi","year":"2017"},{"key":"ref16","first-page":"3570","article-title":"On orthogonality and learning recurrent networks with long term dependencies","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Vorontsov","year":"2017"},{"key":"ref17","first-page":"854","article-title":"Parseval networks: Improving robustness to adversarial examples","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Cisse","year":"2017"},{"key":"ref18","article-title":"Reducing overfitting in deep networks by decorrelating representations","author":"Cogswell","year":"2015"},{"key":"ref19","first-page":"4266","article-title":"Can we gain more from orthogonality regularizations in training deep networks?","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"31","author":"Bansal","year":"2018"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v32i1.11768"},{"key":"ref21","first-page":"521","article-title":"Gradient descent with identity initialization efficiently learns positive definite linear transformations by deep residual networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Bartlett","year":"2018"},{"key":"ref22","article-title":"A convergence analysis of gradient descent for deep linear neural networks","author":"Arora","year":"2018"},{"key":"ref23","article-title":"On the global convergence of training deep linear ResNets","author":"Zou","year":"2020"},{"key":"ref24","first-page":"2691","article-title":"Exponential convergence time of gradient descent for one-dimensional deep linear neural networks","volume-title":"Proc. Conf. Learn. Theory","author":"Shamir","year":"2019"},{"key":"ref25","first-page":"6676","article-title":"On the convergence rate of training recurrent neural networks","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"32","author":"Allen-Zhu","year":"2019"},{"key":"ref26","article-title":"Convergence of gradient descent for deep neural networks","author":"Chatterjee","year":"2022"},{"key":"ref27","first-page":"4577","article-title":"A local convergence theory for mildly over-parameterized two-layer neural network","volume-title":"Proc. Conf. Learn. Theory","author":"Zhou","year":"2021"},{"key":"ref28","first-page":"1524","article-title":"Learning one-hidden-layer relu networks via gradient descent","volume-title":"Proc. 22nd Int. Conf. Artif. Intell. Statist.","author":"Zhang","year":"2019"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1214\/21-aos2061"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1137\/20M1321000"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1007\/s10851-019-00889-w"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.5555\/3045390.3045493"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.2021.3049171"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.123"},{"key":"ref35","first-page":"5393","article-title":"Dynamical isometry and a mean field theory of CNNs: How to train 10,000-layer vanilla convolutional neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Xiao","year":"2018"},{"key":"ref36","first-page":"31697","article-title":"Are all losses created equal: A neural collapse perspective","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"35","author":"Zhou","year":"2022"}],"container-title":["IEEE Signal Processing Letters"],"original-title":[],"link":[{"URL":"https:\/\/ieeexplore.ieee.org\/ielam\/97\/10380231\/10461121-aam.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/97\/10380231\/10461121.pdf?arnumber=10461121","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,4,24]],"date-time":"2024-04-24T17:36:56Z","timestamp":1713980216000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10461121\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024]]},"references-count":36,"URL":"https:\/\/doi.org\/10.1109\/lsp.2024.3374085","relation":{},"ISSN":["1070-9908","1558-2361"],"issn-type":[{"value":"1070-9908","type":"print"},{"value":"1558-2361","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024]]}}}