{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,28]],"date-time":"2026-02-28T18:29:22Z","timestamp":1772303362938,"version":"3.50.1"},"reference-count":41,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"9","license":[{"start":{"date-parts":[[2019,5,1]],"date-time":"2019-05-01T00:00:00Z","timestamp":1556668800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2019,5,1]],"date-time":"2019-05-01T00:00:00Z","timestamp":1556668800000},"content-version":"am","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2019,5,1]],"date-time":"2019-05-01T00:00:00Z","timestamp":1556668800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2019,5,1]],"date-time":"2019-05-01T00:00:00Z","timestamp":1556668800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["1500713"],"award-info":[{"award-number":["1500713"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["1514056"],"award-info":[{"award-number":["1514056"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["1505970"],"award-info":[{"award-number":["1505970"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["1711471"],"award-info":[{"award-number":["1711471"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["U1509215"],"award-info":[{"award-number":["U1509215"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["61621063"],"award-info":[{"award-number":["61621063"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Program for Changjiang Scholars and Innovative Research Team in University","award":["IRT1208"],"award-info":[{"award-number":["IRT1208"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Signal Process."],"published-print":{"date-parts":[[2019,5,1]]},"DOI":"10.1109\/tsp.2019.2904921","type":"journal-article","created":{"date-parts":[[2019,3,20]],"date-time":"2019-03-20T18:43:45Z","timestamp":1553107425000},"page":"2357-2370","source":"Crossref","is-referenced-by-count":114,"title":["Learning ReLU Networks on Linearly Separable Data: Algorithm, Optimality, and Generalization"],"prefix":"10.1109","volume":"67","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7266-2412","authenticated-orcid":false,"given":"Gang","family":"Wang","sequence":"first","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8264-6056","authenticated-orcid":false,"given":"Georgios B.","family":"Giannakis","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2449-9793","authenticated-orcid":false,"given":"Jie","family":"Chen","sequence":"additional","affiliation":[]}],"member":"263","reference":[{"key":"ref39","first-page":"4140","article-title":"Recovery guarantees for one-hidden-layer neural networks","volume":"70","author":"zhong","year":"2017","journal-title":"Proc IEEE Intern Conf on Machine Learning"},{"key":"ref38","article-title":"Real-time power system state estimation and forecasting via deep neural networks","author":"zhang","year":"2018"},{"key":"ref33","article-title":"Very deep convolutional networks for large-scale image recognition","author":"simonyan","year":"2014"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9781107298019"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1037\/h0042519"},{"key":"ref30","first-page":"2847","article-title":"On the expressive power of deep neural networks","volume":"70","author":"raghu","year":"2017","journal-title":"Proc IEEE Intern Conf on Machine Learning"},{"key":"ref37","article-title":"Convergence of SGD in learning ReLU models with separable data","author":"xu","year":"2018"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.2017.2756858"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/72.809097"},{"key":"ref34","first-page":"2007","article-title":"Learning ReLU via gradient descent","author":"soltanolkotabi","year":"2017","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/34.107014"},{"key":"ref40","first-page":"1","article-title":"Critical points of linear neural networks: Analytical forms and landscape properties","author":"zhou","year":"2018","journal-title":"Proc Int Conf on Learning Rep"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/72.105415"},{"key":"ref12","article-title":"Fitting ReLUs via SGD and Quantized SGD","author":"kalan","year":"2019"},{"key":"ref13","first-page":"586","article-title":"Deep learning without poor local minima","author":"kawaguchi","year":"2016","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref14","article-title":"Elimination of all bad local minima in deep learning","author":"kawaguchi","year":"2019"},{"key":"ref15","article-title":"Generalization in deep learning","author":"kawaguchi","year":"2017"},{"key":"ref16","first-page":"1097","article-title":"ImageNet classification with deep convolutional neural networks","author":"krizhevsky","year":"2012","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref17","first-page":"2908","article-title":"Deep linear neural networks with arbitrary loss: All local minima are global","author":"laurent","year":"2018","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref18","first-page":"2914","article-title":"The multilinear structure of ReLU networks","volume":"80","author":"laurent","year":"2018","journal-title":"Proc IEEE Intern Conf on Machine Learning"},{"key":"ref19","article-title":"Over-parameterized deep neural networks have no strict local minima for any continuous activations","author":"li","year":"2018"},{"key":"ref28","article-title":"Towards moderate overparameterization: Global convergence guarantees for training shallow neural networks","author":"oymak","year":"2019"},{"key":"ref4","first-page":"1","article-title":"SGD learns over-parameterized networks that provably generalize on linearly separable data","author":"brutzkus","year":"2018","journal-title":"Proc Int Conf on Learning Rep"},{"key":"ref27","article-title":"Stochastic gradient descent learns state equations with nonlinear activations","author":"oymak","year":"2018"},{"key":"ref3","first-page":"494","article-title":"Training a 3-node neural network is NP-complete","author":"blum","year":"1988","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref6","article-title":"UCI machine learning repository","author":"dheeru","year":"2017"},{"key":"ref29","first-page":"3360","article-title":"Exponential expressivity in deep neural networks through transient chaos","author":"poole","year":"2016","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1137\/1.9781611971309"},{"key":"ref8","first-page":"797","article-title":"Escaping from saddle points—Online stochastic gradient for tensor decomposition","volume":"40","author":"ge","year":"2015","journal-title":"Proc Conf Learn Theory"},{"key":"ref7","article-title":"Computing nonvacuous generalization bounds for deep (stochastic) neural networks with many more parameters than training data","author":"dziugaite","year":"2007"},{"key":"ref2","first-page":"316","article-title":"Exponentially many local minima for single neurons","author":"auer","year":"1995","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref9","volume":"1","author":"goodfellow","year":"2016","journal-title":"Deep Learning"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1996.8.3.643"},{"key":"ref20","article-title":"Adding one neuron can eliminate all bad local minima","author":"liang","year":"2018"},{"key":"ref22","first-page":"807","article-title":"Rectified linear units improve restricted Boltzmann machines","author":"nair","year":"2010","journal-title":"Proc IEEE Intern Conf on Machine Learning"},{"key":"ref21","article-title":"Relating data compression and learnability","author":"littlestone","year":"1986"},{"key":"ref24","article-title":"On connected sublevel sets in deep learning","author":"nguyen","year":"2019"},{"key":"ref41","first-page":"1338","article-title":"Gradient descent\nlearns one-hidden-layer CNN: Don't be afraid of spurious local minima","volume":"80","author":"du","year":"2018","journal-title":"Proc IEEE Intern Conf on Machine Learning"},{"key":"ref23","article-title":"In search of the real inductive bias: On the role of implicit regularization in deep learning","author":"neyshabur","year":"2014"},{"key":"ref26","first-page":"615","article-title":"On convergence proofs for perceptrons","volume":"12","author":"novikoff","year":"0","journal-title":"Proc Symp Math Theory Automata"},{"key":"ref25","first-page":"3727","article-title":"Optimization landscape and expressivity of deep CNNs","author":"nguyen","year":"0","journal-title":"Proc Int Conf Mach Learn"}],"container-title":["IEEE Transactions on Signal Processing"],"original-title":[],"link":[{"URL":"https:\/\/ieeexplore.ieee.org\/ielaam\/78\/8673709\/8671751-aam.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/78\/8673709\/08671751.pdf?arnumber=8671751","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,7,13]],"date-time":"2022-07-13T20:49:39Z","timestamp":1657745379000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/8671751\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2019,5,1]]},"references-count":41,"journal-issue":{"issue":"9"},"URL":"https:\/\/doi.org\/10.1109\/tsp.2019.2904921","relation":{},"ISSN":["1053-587X","1941-0476"],"issn-type":[{"value":"1053-587X","type":"print"},{"value":"1941-0476","type":"electronic"}],"subject":[],"published":{"date-parts":[[2019,5,1]]}}}