{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,25]],"date-time":"2025-11-25T06:07:38Z","timestamp":1764050858724,"version":"3.45.0"},"reference-count":43,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,10,9]],"date-time":"2025-10-09T00:00:00Z","timestamp":1759968000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,9]],"date-time":"2025-10-09T00:00:00Z","timestamp":1759968000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,9]]},"DOI":"10.1109\/dsaa65442.2025.11247980","type":"proceedings-article","created":{"date-parts":[[2025,11,24]],"date-time":"2025-11-24T18:56:45Z","timestamp":1764010605000},"page":"1-8","source":"Crossref","is-referenced-by-count":0,"title":["Optimal Utility Bounds for Differentially Private Gradient Descent in Three-Layer Neural Networks"],"prefix":"10.1109","author":[{"given":"Puyu","family":"Wang","sequence":"first","affiliation":[{"name":"RPTU Kaiserslautern-Landau,Department of Computer Science,Kaiserslautern,Germany"}]},{"given":"Yunwen","family":"Lei","sequence":"additional","affiliation":[{"name":"The University of Hong Kong,Department of Mathematics,Hong Kong,China"}]},{"given":"Marius","family":"Kloft","sequence":"additional","affiliation":[{"name":"RPTU Kaiserslautern-Landau,Department of Computer Science,Kaiserslautern,Germany"}]},{"given":"Yiming","family":"Ying","sequence":"additional","affiliation":[{"name":"School of Mathematics and Statistics, The University of Sydney,Sydney,Australia"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1145\/3065386"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/MSP.2012.2205597"},{"journal-title":"Neural machine translation by jointly learning to align and translate","year":"2014","author":"Bahdanau","key":"ref3"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.13140\/RG.2.2.18893.74727"},{"key":"ref5","first-page":"242","article-title":"A convergence theory for deep learning via over-parameterization","volume-title":"International Conference on Machine Learning","author":"Allen-Zhu","year":"2019"},{"key":"ref6","first-page":"3764","article-title":"Convergence of cubic regularization for nonconvex optimization under kl property","author":"Zhou","year":"2018","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref7","first-page":"322","article-title":"Fine-grained analysis of optimization and generalization for overparameterized two-layer neural networks","volume-title":"International Conference on Machine Learning. PMLR","author":"Arora","year":"2019"},{"key":"ref8","article-title":"Stability and generalization analysis of gradient methods for shallow neural networks","volume":"35","author":"Lei","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/SP.2017.41"},{"key":"ref10","article-title":"Deep leakage from gradients","volume":"32","author":"Zhu","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1007\/11681878_14"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1561\/0400000042"},{"key":"ref13","article-title":"Stability of stochastic gradient descent on non smooth convex losses","volume":"33","author":"Bassily","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/FOCS.2014.56"},{"key":"ref15","first-page":"9317","article-title":"Differentially private stochas-tic optimization: New results in convex and non-convex settings","volume":"34","author":"Bassily","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/3357713.3384335"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2024.127557"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1016\/j.acha.2021.09.001"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2021\/446"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1145\/2976749.2978318"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1162\/99608f92.cfc5dd25"},{"key":"ref22","article-title":"On the convergence and calibration of deep learning with differential privacy","volume":"2023","author":"Bu","year":"2023","journal-title":"Transactions on machine learning research"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1145\/2810103.2813687"},{"key":"ref24","first-page":"27131","article-title":"Deep learning with label differential privacy","volume":"34","author":"Ghazi","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v35i10.17123"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/ICDM.2017.48"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-26412-2_37"},{"issue":"3","key":"ref28","article-title":"Differentially private empirical risk minimization","volume":"12","author":"Chaudhuri","year":"2011","journal-title":"Journal of Machine Learning Research"},{"key":"ref29","first-page":"1990","article-title":"Learning with gradient descent and weakly convex losses","volume-title":"International Conference on Artificial Intelligence and Statistics. PMLR","author":"Richards","year":"2021"},{"key":"ref30","article-title":"On exact computation with an infinitely wide neural net","volume":"32","author":"Arora","year":"2019","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref31","article-title":"Neural tangent kernel: Con-vergence and generalization in neural networks","volume":"31","author":"Jacot","year":"2018","journal-title":"Advances in neural information processing systems"},{"key":"ref32","article-title":"Gradient descent provably optimizes over-parameterized neural networks","author":"Du","year":"2018","journal-title":"International Conference on Learning Representations"},{"key":"ref33","first-page":"1675","article-title":"Gradient descent finds global minima of deep neural networks","volume-title":"International conference on machine learning. PMLR","author":"Du","year":"2019"},{"key":"ref34","article-title":"On the global convergence of gradient descent for over-parameterized models using optimal transport","volume":"31","author":"Chizat","year":"2018","journal-title":"Advances in neural information processing systems"},{"key":"ref35","article-title":"On lazy training in differentiable programming","volume":"32","author":"Chizat","year":"2019","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref36","first-page":"2388","article-title":"Mean-field theory of two-layers neural networks: dimension-free bounds and kernel limit","volume-title":"Conference on Learning Theory. PMLR","author":"Mei","year":"2019"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1806579115"},{"key":"ref38","first-page":"11279","article-title":"Private stochastic convex optimization with optimal rates","author":"Bassily","year":"2019","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/CSF.2017.11"},{"key":"ref40","article-title":"Stability & generalisation of gradient descent for shallow neural networks without the neural tangent kernel","volume":"34","author":"Richards","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1162\/neco_a_01725"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2021.07.001"},{"key":"ref43","first-page":"2199","article-title":"Smoothness, low noise and fast rates","author":"Srebro","year":"2010","journal-title":"Advances in Neural Information Processing Systems"}],"event":{"name":"2025 IEEE 12th International Conference on Data Science and Advanced Analytics (DSAA)","start":{"date-parts":[[2025,10,9]]},"location":"Birmingham, United Kingdom","end":{"date-parts":[[2025,10,12]]}},"container-title":["2025 IEEE 12th International Conference on Data Science and Advanced Analytics (DSAA)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11247920\/11247921\/11247980.pdf?arnumber=11247980","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,25]],"date-time":"2025-11-25T05:59:20Z","timestamp":1764050360000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11247980\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,9]]},"references-count":43,"URL":"https:\/\/doi.org\/10.1109\/dsaa65442.2025.11247980","relation":{},"subject":[],"published":{"date-parts":[[2025,10,9]]}}}