{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,6]],"date-time":"2026-05-06T07:22:28Z","timestamp":1778052148901,"version":"3.51.4"},"reference-count":50,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"NSFC","doi-asserted-by":"crossref","award":["12471096"],"award-info":[{"award-number":["12471096"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]},{"name":"National Key Research and Development Program of China","award":["2021YFA1003500"],"award-info":[{"award-number":["2021YFA1003500"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Signal Process."],"published-print":{"date-parts":[[2026]]},"DOI":"10.1109\/tsp.2026.3676735","type":"journal-article","created":{"date-parts":[[2026,3,23]],"date-time":"2026-03-23T20:11:48Z","timestamp":1774296708000},"page":"1491-1507","source":"Crossref","is-referenced-by-count":0,"title":["Convergence Theory of Sharpness-Aware Minimization With Interpolating Neural Networks"],"prefix":"10.1109","volume":"74","author":[{"ORCID":"https:\/\/orcid.org\/0009-0005-6640-7684","authenticated-orcid":false,"given":"Tan","family":"Sun","sequence":"first","affiliation":[{"name":"Center for Data Science, Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9418-3095","authenticated-orcid":false,"given":"Yusu","family":"Hong","sequence":"additional","affiliation":[{"name":"Center for Data Science, Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4507-9424","authenticated-orcid":false,"given":"Junhong","family":"Lin","sequence":"additional","affiliation":[{"name":"Center for Data Science, Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1214\/aoms\/1177729586"},{"key":"ref2","first-page":"2874","article-title":"On large-batch training for deep learning: Generalization gap and sharp minima","volume-title":"Proc. Int. Conf. Learn. Representations","volume":"3","author":"Keskar","year":"2017"},{"key":"ref3","first-page":"5947","article-title":"Exploring generalization in deep learning","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"30","author":"Neyshabur","year":"2017"},{"key":"ref4","first-page":"2315","article-title":"Sharpness-aware minimization for efficiently improving generalization","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Foret","year":"2021"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.acl-long.508"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.findings-emnlp.300"},{"key":"ref7","first-page":"5905","article-title":"ASAM: Adaptive sharpness-aware minimization for scale-invariant learning of deep neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Kwon","year":"2021"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2023.10.044"},{"key":"ref9","first-page":"1","article-title":"$\\nu$SAM: Memory-efficient sharpness-aware minimization via nuclear norm constraints","author":"Pethick","year":"2025","journal-title":"Trans. Mach. Learn. Res."},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2244"},{"key":"ref11","first-page":"2553","article-title":"Surrogate gap minimization improves sharpness-aware training","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Zhuang","year":"2022"},{"key":"ref12","first-page":"639","article-title":"Towards understanding sharpness-aware minimization","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Andriushchenko","year":"2022"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1140"},{"key":"ref14","first-page":"70861","article-title":"Enhancing sharpness-aware optimization through variance suppression","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"36","author":"Li","year":"2023"},{"key":"ref15","article-title":"The effects of overparameterization on sharpness-aware minimization: An empirical and theoretical analysis","author":"Shin","year":"2023"},{"key":"ref16","first-page":"51333","article-title":"SAMPa: Sharpness-aware minimization parallelized","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"37","author":"Xie","year":"2026"},{"key":"ref17","article-title":"On saddle point avoidance and stationary distribution of sharpness-aware minimization","volume-title":"Preprints","author":"Sun","year":"2018"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01204"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1703"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00538"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP49660.2025.10890114"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0419"},{"key":"ref23","first-page":"15149","article-title":"Sharpness-aware minimization: General analysis and improved rates","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Oikonomou","year":"2025"},{"key":"ref24","first-page":"1128","article-title":"Global optimality conditions for deep neural networks","volume-title":"Proc. Int. Conf. Learn. Representations","volume":"2","author":"Yun","year":"2018"},{"key":"ref25","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":"ref26","first-page":"242","article-title":"A convergence theory for deep learning via over-parameterization","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Allen-Zhu","year":"2019"},{"key":"ref27","first-page":"5403","article-title":"Gradient descent provably optimizes over-parameterized neural networks","volume-title":"Proc. Int. Conf. Learn. Representations","volume":"7","author":"Du","year":"2019"},{"key":"ref28","first-page":"1675","article-title":"Gradient descent finds global minima of deep neural networks","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Du","year":"2019"},{"key":"ref29","first-page":"2055","article-title":"An improved analysis of training over-parameterized deep neural networks","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"32","author":"Zou","year":"2019"},{"key":"ref30","first-page":"3267","article-title":"A convergence analysis of gradient descent for deep linear neural networks","volume-title":"Proc. Int. Conf. Learn. Representations","volume":"5","author":"Arora","year":"2019"},{"key":"ref31","first-page":"4689","article-title":"On the global convergence of training deep linear ResNets","volume-title":"Proc. Int. Conf. Learn. Representations","volume":"6","author":"Zou","year":"2020"},{"key":"ref32","first-page":"4239","article-title":"Polylogarithmic width suffices for gradient descent to achieve arbitrarily small test error with shallow ReLU networks","volume-title":"Proc. Int. Conf. Learn. Representations","volume":"6","author":"Ji","year":"2020"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1214\/20-AOS1945"},{"key":"ref34","first-page":"8157","article-title":"Learning overparameterized neural networks via stochastic gradient descent on structured data","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"31","author":"Li","year":"2018"},{"key":"ref35","first-page":"322","article-title":"Fine-grained analysis of optimization and generalization for overparameterized two-layer neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Arora","year":"2019"},{"key":"ref36","article-title":"Gradient descent can learn less over-parameterized two-layer neural networks on classification problems","author":"Nitanda","year":"2019"},{"issue":"156","key":"ref37","first-page":"1","article-title":"Generalization and stability of interpolating neural networks with minimal width","volume":"25","author":"Taheri","year":"2024","journal-title":"J. Mach. Learn. Res."},{"key":"ref38","first-page":"1049","article-title":"On the design of loss functions for classification: Theory, robustness to outliers, and SavageBoost","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"21","author":"Masnadi-Shirazi","year":"2008"},{"key":"ref39","first-page":"41032","article-title":"Sharper guarantees for learning neural network classifiers with gradient methods","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Taheri","year":"2025"},{"key":"ref40","first-page":"8571","article-title":"Neural tangent kernel: Convergence and generalization in neural networks","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"31","author":"Jacot","year":"2018"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i04.5736"},{"key":"ref42","first-page":"6796","article-title":"How much over-parameterization is sufficient to learn deep ReLU networks?","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Chen","year":"2021"},{"key":"ref43","first-page":"15954","article-title":"On the linearity of large non-linear models: When and why the tangent kernel is constant","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"33","author":"Liu","year":"2020"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1016\/j.acha.2021.12.009"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.123"},{"key":"ref46","first-page":"9712","article-title":"Regularization matters: Generalization and optimization of neural nets vs their induced kernel","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"32","author":"Wei","year":"2019"},{"key":"ref47","doi-asserted-by":"crossref","DOI":"10.1109\/5.726791","article-title":"Gradient-based learning applied to document recognition","author":"Lecun","year":"1998"},{"key":"ref48","article-title":"Learning multiple layers of features from tiny images","author":"Krizhevsky","year":"2009"},{"key":"ref49","article-title":"Introduction to the non-asymptotic analysis of random matrices","author":"Vershynin","year":"2011"},{"key":"ref50","article-title":"Optimization analysis for two-layer neural networks trained with AdaGrad-Norm: interpolating models and minimal width","author":"Hong","year":"2025"}],"container-title":["IEEE Transactions on Signal Processing"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/78\/11345506\/11454682.pdf?arnumber=11454682","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,6]],"date-time":"2026-05-06T06:45:58Z","timestamp":1778049958000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11454682\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":50,"URL":"https:\/\/doi.org\/10.1109\/tsp.2026.3676735","relation":{},"ISSN":["1053-587X","1941-0476"],"issn-type":[{"value":"1053-587X","type":"print"},{"value":"1941-0476","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026]]}}}