{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,13]],"date-time":"2026-05-13T18:21:36Z","timestamp":1778696496008,"version":"3.51.4"},"reference-count":34,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"2","license":[{"start":{"date-parts":[[2020,10,1]],"date-time":"2020-10-01T00:00:00Z","timestamp":1601510400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2020,10,1]],"date-time":"2020-10-01T00:00:00Z","timestamp":1601510400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2020,10,1]],"date-time":"2020-10-01T00:00:00Z","timestamp":1601510400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/100000006","name":"Office of Naval Research","doi-asserted-by":"publisher","award":["N00014-13-1-0338"],"award-info":[{"award-number":["N00014-13-1-0338"]}],"id":[{"id":"10.13039\/100000006","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Princeton University by iFlytek"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Artif. Intell."],"published-print":{"date-parts":[[2020,10]]},"DOI":"10.1109\/tai.2021.3051357","type":"journal-article","created":{"date-parts":[[2021,6,30]],"date-time":"2021-06-30T15:33:57Z","timestamp":1625067237000},"page":"121-129","source":"Crossref","is-referenced-by-count":39,"title":["Can Shallow Neural Networks Beat the Curse of Dimensionality? A Mean Field Training Perspective"],"prefix":"10.1109","volume":"1","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3766-5332","authenticated-orcid":false,"given":"Stephan","family":"Wojtowytsch","sequence":"first","affiliation":[{"name":"Program of Applied and Computational Mathematics, Princeton University, Princetion, NJ, USA"}]},{"given":"Weinan","family":"E","sequence":"additional","affiliation":[{"name":"Department of Mathematics and the Program of Applied and Computational Mathematics, Princeton University, Princetion, NJ, USA"}]}],"member":"263","reference":[{"key":"ref33","article-title":"CoD test simulations","author":"wojtowytsch","year":"2020"},{"key":"ref32","article-title":"Optimal Transport: Old and New","volume":"338","author":"villani","year":"2008"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1137\/18M1192184"},{"key":"ref30","first-page":"1135","article-title":"Distribution-specific hardness of learning neural networks","volume":"19","author":"shamir","year":"2018","journal-title":"J Mach Learn Res"},{"key":"ref34","article-title":"On the global convergence of gradient descent training for two-layer ReLU networks in the mean field regime","author":"wojtowytsch","year":"2020"},{"key":"ref10","article-title":"On lazy training in differentiable programming","author":"chizat","year":"2018"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1007\/BF02551274"},{"key":"ref12","article-title":"Gradient descent provably optimizes over-parameterized neural networks","author":"du","year":"2018"},{"key":"ref13","article-title":"Barron spaces and the compositional function spaces for neural network models","author":"weinan","year":"2019"},{"key":"ref14","article-title":"Machine learning from a continuous viewpoint","author":"weinan","year":"2019"},{"key":"ref15","doi-asserted-by":"crossref","first-page":"1407","DOI":"10.4310\/CMS.2019.v17.n5.a11","article-title":"A. priori estimates of the population risk for two-layer neural networks","volume":"17","author":"weinan","year":"2019","journal-title":"Comm Math Sci"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1007\/s11425-019-1628-5"},{"key":"ref17","article-title":"Kolmogorov width decay and poor approximators in machine learning: Shallow neural networks, random feature models and neural tangent kernels","author":"weinan","year":"2020"},{"key":"ref18","article-title":"A priori estimates for classification problems using neural networks","author":"weinan","year":"2020"},{"key":"ref19","article-title":"Representation formulas and pointwise properties for barron functions","author":"weinan","year":"2020"},{"key":"ref28","article-title":"Neural networks as interacting particle systems: asymptotic convexity of the loss landscape and universal scaling of the approximation error","author":"rotskoff","year":"2018"},{"key":"ref4","first-page":"629","article-title":"Breaking the curse of dimensionality with convex neural networks","volume":"18","author":"bach","year":"2017","journal-title":"J Mach Learn Res"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1145\/3186563"},{"key":"ref3","first-page":"6481","article-title":"Maximum mean discrepancy gradient flow","author":"arbel","year":"0","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1016\/j.crma.2018.03.003"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-20828-2"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/18.256500"},{"key":"ref8","first-page":"3036","article-title":"On the global convergence of gradient descent for over-parameterized models using optimal transport","author":"chizat","year":"0","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref7","article-title":"A generalized neural tangent kernel analysis for two-layer neural networks","volume":"33","author":"chen","year":"2020","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref2","author":"ambrosio","year":"2008","journal-title":"Gradient Flows in Metric Spaces and in the Space of Probability Measures"},{"key":"ref9","article-title":"Implicit bias of gradient descent for wide two-layer neural networks trained with the logistic loss","author":"chizat","year":"2020"},{"key":"ref1","article-title":"Provable limitations of deep learning","author":"abbe","year":"2018"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1016\/0893-6080(91)90009-T"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.4310\/AMSA.2019.v4.n1.a1"},{"key":"ref21","article-title":"Mean-Field Langevin dynamics and energy landscape of neural networks","author":"hu","year":"2019"},{"key":"ref24","first-page":"855","article-title":"On the computational efficiency of training neural networks","author":"livni","year":"0","journal-title":"Adv Neural Inf Process Syst"},{"key":"ref23","article-title":"Dynamics of stochastic gradient algorithms","author":"li","year":"2015"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1806579115"},{"key":"ref25","article-title":"Mean-field theory of two-layers neural networks: Dimension-free bounds and kernel limit","author":"mei","year":"2019"}],"container-title":["IEEE Transactions on Artificial Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/9078688\/9368888\/09321497.pdf?arnumber=9321497","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,8,23]],"date-time":"2025-08-23T01:09:47Z","timestamp":1755911387000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9321497\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,10]]},"references-count":34,"journal-issue":{"issue":"2"},"URL":"https:\/\/doi.org\/10.1109\/tai.2021.3051357","relation":{},"ISSN":["2691-4581"],"issn-type":[{"value":"2691-4581","type":"electronic"}],"subject":[],"published":{"date-parts":[[2020,10]]}}}