{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,16]],"date-time":"2026-04-16T21:21:38Z","timestamp":1776374498548,"version":"3.51.2"},"reference-count":52,"publisher":"Oxford University Press (OUP)","issue":"12","license":[{"start":{"date-parts":[[2024,9,20]],"date-time":"2024-09-20T00:00:00Z","timestamp":1726790400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/pages\/standard-publication-reuse-rights"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key Research and Development Program of China","doi-asserted-by":"publisher","award":["2022YFB2703301"],"award-info":[{"award-number":["2022YFB2703301"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,12,20]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Federated learning (FL) as a privacy-preserving technology enables multiple clients to collaboratively train models on decentralized data. However, transmitting model parameters between local clients and the central server can potentially result in information leakage. Differentially private federated learning (DPFL) has emerged as a promising solution to enhance privacy. Nevertheless, existing DPFL schemes suffer from two issues: (i) most schemes that aim to achieve desired model accuracy may incur a high privacy budget. (ii) several schemes that consider the trade-off between privacy and accuracy by utilizing linear clipping bound may distort numerous model parameters. In this paper, we first propose FDP-FL, a flexible differential privacy approach for FL. FDP-FL introduces a novel series sum privacy budget\u00a0allocation instead of uniform allocation and enables adaptive and nonlinear noise scale decay. In this way, a tight bound for cumulative privacy loss can be achieved while optimizing model accuracy. Then in order to mitigate gradient leakages caused by honest-but-curious clients and server, we further design client-level FDP-FL and record-level FDP-FL, respectively. Experimental results demonstrate that our FDP-FL improves model accuracy by $\\sim $13.3% compared with the basic DP-FL under a fixed privacy budget and outperforms existing trade-off schemes with the same hyperparameter setting.<\/jats:p>","DOI":"10.1093\/comjnl\/bxae081","type":"journal-article","created":{"date-parts":[[2024,9,21]],"date-time":"2024-09-21T05:08:14Z","timestamp":1726895294000},"page":"3180-3195","source":"Crossref","is-referenced-by-count":4,"title":["FDP-FL: differentially private federated learning with flexible privacy budget\u00a0allocation"],"prefix":"10.1093","volume":"67","author":[{"given":"Wenjun","family":"Qian","sequence":"first","affiliation":[{"name":"School of Software and Microelectronics , Peking University, No. 24 Jinyuan Road, Beijing 102600,","place":["China"]},{"name":"National Engineering Research Center for Software Engineering , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"PKU-OCTA Laboratory for Blockchain and Privacy Computing , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Qingni","family":"Shen","sequence":"additional","affiliation":[{"name":"School of Software and Microelectronics , Peking University, No. 24 Jinyuan Road, Beijing 102600,","place":["China"]},{"name":"National Engineering Research Center for Software Engineering , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"PKU-OCTA Laboratory for Blockchain and Privacy Computing , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xiaoyi","family":"Chen","sequence":"additional","affiliation":[{"name":"National Engineering Research Center for Software Engineering , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"PKU-OCTA Laboratory for Blockchain and Privacy Computing , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"School of Computer Science , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Cong","family":"Li","sequence":"additional","affiliation":[{"name":"National Engineering Research Center for Software Engineering , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"PKU-OCTA Laboratory for Blockchain and Privacy Computing , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"School of Computer Science , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yuejian","family":"Fang","sequence":"additional","affiliation":[{"name":"School of Software and Microelectronics , Peking University, No. 24 Jinyuan Road, Beijing 102600,","place":["China"]},{"name":"National Engineering Research Center for Software Engineering , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"PKU-OCTA Laboratory for Blockchain and Privacy Computing , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhonghai","family":"Wu","sequence":"additional","affiliation":[{"name":"School of Software and Microelectronics , Peking University, No. 24 Jinyuan Road, Beijing 102600,","place":["China"]},{"name":"National Engineering Research Center for Software Engineering , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"PKU-OCTA Laboratory for Blockchain and Privacy Computing , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]},{"name":"School of Computer Science , Peking University, No. 5 Yiheyuan Road, Beijing 100871,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2024,9,20]]},"reference":[{"key":"2025010523430736800_ref1","first-page":"1273","article-title":"Communication-efficient learning of deep networks from decentralized data","volume-title":"Proc. 20th int. conf. Artif. Intell. Stat., Fort Lauderdale, FL, USA, 20\u201322 April","author":"McMahan","year":"2017"},{"key":"2025010523430736800_ref2","doi-asserted-by":"publisher","first-page":"107818","DOI":"10.1016\/j.compeleceng.2022.107818","article-title":"A review on federated learning towards image processing","volume":"99","author":"KhoKhar","year":"2022","journal-title":"Comput Electr Eng"},{"key":"2025010523430736800_ref3","doi-asserted-by":"publisher","first-page":"66:1","DOI":"10.1145\/3510033","article-title":"Fedbert: When federated learning meets pre-training","volume":"13","author":"Tian","year":"2022","journal-title":"ACM Trans Intell Syst Technol"},{"key":"2025010523430736800_ref4","first-page":"3080","article-title":"Training speech recognition models with federated learning: a quality\/cost framework","volume-title":"Proc. 46th IEEE int. conf. Acoust., speech sig. Process., Toronto, ON, Canada, 6\u201311 June","author":"Guliani","year":"2021"},{"key":"2025010523430736800_ref5","doi-asserted-by":"publisher","first-page":"275","DOI":"10.3390\/jpm12020275","article-title":"A comprehensive analysis of recent deep and federated-learning-based methodologies for brain tumor diagnosis","volume":"12","author":"Naeem","year":"2022","journal-title":"J Pers Med"},{"key":"2025010523430736800_ref6","first-page":"739","article-title":"Comprehensive privacy analysis of deep learning: passive and active white-box inference attacks against centralized and federated learning","volume-title":"Proc. 40th IEEE Symp. Secur. Priv., San Francisco, CA, USA, 19\u201323 may","author":"Nasr","year":"2019"},{"key":"2025010523430736800_ref7","first-page":"691","article-title":"Exploiting unintended feature leakage in collaborative learning","volume-title":"Proc. 40th IEEE Symp. Secur. Priv., San Francisco, CA, USA, 19\u201323 may","author":"Melis","year":"2019"},{"key":"2025010523430736800_ref8","first-page":"1","article-title":"Gan enhanced membership inference: A passive local attack in federated learning","volume-title":"Proc. 2020 IEEE int. conf. Commun., Dublin, Ireland, 7\u201311 June","author":"Zhang","year":"2020"},{"key":"2025010523430736800_ref9","first-page":"603","article-title":"Deep models under the Gan: Information leakage from collaborative deep learning","volume-title":"Proc. 24th ACM conf. Comput. Commun. Secur., Dallas, TX, USA, October 30\u2013November 03","author":"Hitaj","year":"2017"},{"key":"2025010523430736800_ref10","first-page":"545","article-title":"A framework for evaluating client privacy leakages in federated learning","volume-title":"Proc. 25th Eur. Symp. Res. Comput. Secur., Guildford, UK, 14\u201318 September","author":"Wei","year":"2020"},{"key":"2025010523430736800_ref11","doi-asserted-by":"crossref","first-page":"2265","DOI":"10.1109\/JIOT.2020.3028110","article-title":"Exploiting unintended property leakage in blockchain-assisted federated learning for intelligent edge computing","volume":"8","author":"Shen","year":"2020","journal-title":"IEEE Internet Things J."},{"key":"2025010523430736800_ref12","article-title":"Differential privacy: a survey of results. Proc. 5th int. conf. Theory appl. Models Comput., Xi\u2019an, China, 25-29 April, pp. 1\u201319. Berlin: Springer","author":"Dwork","year":"2008"},{"key":"2025010523430736800_ref13","article-title":"Differentially private federated learning: a client level perspective","author":"Geyer","year":"2017"},{"key":"2025010523430736800_ref14","first-page":"1","article-title":"Learning differentially private recurrent language models","volume-title":"Proc. 6th int. conf. Learn. Represent., Vancouver, BC, Canada, April 30\u2013May 3","author":"McMahan","year":"2018"},{"key":"2025010523430736800_ref15","doi-asserted-by":"publisher","first-page":"3454","DOI":"10.1109\/TIFS.2020.2988575","article-title":"Federated learning with differential privacy: Algorithms and performance analysis","volume":"15","author":"Wei","year":"2020","journal-title":"IEEE Trans Inf Forensics Secur"},{"key":"2025010523430736800_ref16","first-page":"61","article-title":"Ldp-fed: federated learning with local differential privacy","volume-title":"Proc. 3rd ACM int. workshop edge syst., anal. Netw., Heraklion, Crete, Greece, 27 April","author":"Truex","year":"2020"},{"key":"2025010523430736800_ref17","first-page":"1","article-title":"Utility optimization of federated learning with differential privacy","volume":"2021","author":"Zhao","year":"2021","journal-title":"Discrete Dyn Nat Soc"},{"key":"2025010523430736800_ref18","first-page":"265","article-title":"Calibrating noise to sensitivity in private data analysis","volume-title":"Proc. 3rd theory crypt. Conf., New York, NY, USA, 4\u20137 march","author":"Dwork","year":"2006"},{"key":"2025010523430736800_ref19","first-page":"308","article-title":"Deep learning with differential privacy","volume-title":"Proc. 23rd ACM conf. Comput. Commun. Secur., Vienna, Austria, 24\u201328 October","author":"Abadi","year":"2016"},{"key":"2025010523430736800_ref20","first-page":"920","article-title":"A better bound gives a hundred rounds: enhanced privacy guarantees via f-divergences","volume-title":"Proc. IEEE int. Symp. Inf. Theor, Los Angeles, CA, USA, 21\u201326 June","author":"Asoodeh","year":"2020"},{"key":"2025010523430736800_ref21","first-page":"2650","article-title":"Federated learning with local differential privacy: trade-offs between privacy, utility, and communication","volume-title":"Proc. 46th IEEE int. conf. Acoust., speech sig. Process., Toronto, ON, Canada, 6\u201311 June","author":"Kim","year":"2021"},{"key":"2025010523430736800_ref22","first-page":"51","article-title":"Boosting and differential privacy","volume-title":"Proc. 51st Annu. IEEE Symp. Found. Comput. Sci., Las Vegas, Nevada, USA, 23\u201326 October","author":"Dwork","year":"2010"},{"key":"2025010523430736800_ref23","first-page":"19","article-title":"Privacy integrated queries: an extensible platform for privacy-preserving data analysis","volume-title":"Proc. 35th ACM int. conf. Manage. Data, Providence, Rhode Island, USA, June 29\u2013July 2","author":"McSherry","year":"2009"},{"key":"2025010523430736800_ref24","first-page":"656","article-title":"Adap DP-FL: Differentially private federated learning with adaptive noise","volume-title":"Proc. 21st IEEE Int. Conf. Trust, Secur. Priv. Comput. Commun., Wuhan, China, 9\u201311 December","author":"Fu","year":"2022"},{"key":"2025010523430736800_ref25","doi-asserted-by":"publisher","first-page":"1884","DOI":"10.1109\/TIFS.2023.3258255","article-title":"Amplitude-varying perturbation for balancing privacy and utility in federated learning","volume":"18","author":"Yuan","year":"2023","journal-title":"IEEE Trans Inf Forensics Secur"},{"key":"2025010523430736800_ref26","first-page":"181","article-title":"Privacy accounting and quality control in the sage differentially private ml platform","volume-title":"Proc. 27th ACM Symp. Oper. Syst. Princ., Huntsville, ON, Canada, 27\u201330 October","author":"L\u00e9cuyer","year":"2019"},{"key":"2025010523430736800_ref27","first-page":"1","article-title":"Privacy budgeting for growing machine learning datasets","volume-title":"Proc. 40th IEEE conf. Comput. Commun., Vancouver, BC, Canada, 10\u201313 May","author":"Li","year":"2021"},{"key":"2025010523430736800_ref28","first-page":"797","article-title":"Gradient-leakage resilient federated learning","volume-title":"Proc. 41st IEEE Int. Conf. Distrib. Comput. Syst., Washington DC, USA, 7\u201310 July","author":"Wei","year":"2021"},{"key":"2025010523430736800_ref29","first-page":"332","article-title":"Differentially private model publishing for deep learning","volume-title":"Proc. 40th IEEE Symp. Secur. Priv., San Francisco, CA, USA, 19\u201323 May","author":"Yu","year":"2019"},{"key":"2025010523430736800_ref30","first-page":"26048","article-title":"Understanding clipping for federated learning: convergence and client-level differential privacy","volume-title":"Proc. 39th Int. Conf. Mach. Learn., Baltimore, Maryland, USA, 17\u201323 July","author":"Zhang","year":"2022"},{"key":"2025010523430736800_ref31","first-page":"10112","article-title":"Differentially private federated learning with local regularization and sparsification","volume-title":"Proc. IEEE Comput. Soc. Conf. Comput. Vision Pattern Recognit., New Orleans, LA, USA, 18\u201324 June","author":"Cheng","year":"2022"},{"key":"2025010523430736800_ref32","doi-asserted-by":"publisher","first-page":"1622","DOI":"10.1109\/COMST.2021.3075439","article-title":"Federated learning for internet of things: a comprehensive survey","volume":"23","author":"Nguyen","year":"2021","journal-title":"IEEE Commun Surv Tutorials"},{"key":"2025010523430736800_ref33","doi-asserted-by":"publisher","first-page":"1205","DOI":"10.1109\/JSAC.2019.2904348","article-title":"Adaptive federated learning in resource constrained edge computing systems","volume":"37","author":"Wang","year":"2019","journal-title":"IEEE J Sel Areas Commun"},{"key":"2025010523430736800_ref34","doi-asserted-by":"publisher","DOI":"10.1016\/j.compeleceng.2023.108622","article-title":"Dynamic context management in context-aware recommender systems","volume":"107","author":"Ali","year":"2023","journal-title":"Comput Electr Eng"},{"key":"2025010523430736800_ref35","doi-asserted-by":"publisher","first-page":"1016","DOI":"10.1093\/comjnl\/bxab025","article-title":"A federated learning approach for privacy protection in context-aware recommender systems","volume":"64","author":"Ali","year":"2021","journal-title":"Comput J"},{"key":"2025010523430736800_ref36","doi-asserted-by":"publisher","first-page":"3347","DOI":"10.1109\/TKDE.2021.3124599","article-title":"A survey on federated learning systems: Vision, hype and reality for data privacy and protection","volume":"35","author":"Li","year":"2023","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"2025010523430736800_ref37","first-page":"6346","article-title":"Distributed learning without distress: privacy-preserving empirical risk minimization","volume-title":"Proc. 32nd Annu. Conf. Neural Inf. Process. Syst., Montreal, QC, Canada, 2\u20138 December","author":"Jayaraman","year":"2018"},{"key":"2025010523430736800_ref38","article-title":"Protection against reconstruction and its applications in private federated learning","author":"Bhowmick","year":"2018"},{"key":"2025010523430736800_ref39","doi-asserted-by":"publisher","first-page":"22204","DOI":"10.1109\/JIOT.2021.3101991","article-title":"Differential privacy meets federated learning under communication constraints","volume":"9","author":"Mohammadi","year":"2022","journal-title":"IEEE Internet Things J"},{"key":"2025010523430736800_ref40","first-page":"4872","article-title":"Personalized privacy-preserving federated learning: optimized trade-off between utility and privacy","volume-title":"Proc. IEEE Global Commun. Conf., Rio de Janeiro, Brazil, 4\u20138 December","author":"Zhou","year":"2022"},{"key":"2025010523430736800_ref41","first-page":"94","article-title":"Mechanism design via differential privacy","volume-title":"Proc. 48th Annu. IEEE Symp. Found. Comput. Sci., Providence, RI, USA, 20\u201323 October","author":"McSherry","year":"2007"},{"key":"2025010523430736800_ref42","doi-asserted-by":"crossref","first-page":"211","DOI":"10.1561\/0400000042","article-title":"The algorithmic foundations of differential privacy","volume":"9","author":"Dwork","year":"2014","journal-title":"Found Trends Theor Comput Sci"},{"key":"2025010523430736800_ref43","first-page":"14747","article-title":"Deep leakage from gradients","volume-title":"Proc. 33rd Annu. Conf. Neural Inf. Process. Syst., Vancouver, BC, Canada, 8\u201314 December","author":"Zhu","year":"2019"},{"key":"2025010523430736800_ref44","first-page":"16937","article-title":"Inverting gradients - how easy is it to break privacy in federated learning?","volume-title":"Proc. 34th Annu. Conf. Neural Inf. Process. Syst., Virtual, 6\u201312 December","author":"Geiping","year":"2020"},{"key":"2025010523430736800_ref45","first-page":"155","article-title":"Privtree: a differentially private algorithm for hierarchical decompositions","volume-title":"Proc. 42rd Int. Conf. Manage. Data, San Francisco, CA, USA, June 26\u2013July 01","author":"Zhang","year":"2016"},{"key":"2025010523430736800_ref46","first-page":"75","article-title":"Smooth sensitivity and sampling in private data analysis","volume-title":"Proc. 39th Annu. ACM Symp. Theor. Comput., San Diego, CA, USA, 11\u201313 June","author":"Nissim","year":"2007"},{"key":"2025010523430736800_ref47","first-page":"17455","article-title":"Differentially private learning with adaptive clipping","volume-title":"Proc. 35th Annu. Conf. Neural Inf. Process. Syst., Virtual, 6\u201314 December","author":"Andrew","year":"2021"},{"key":"2025010523430736800_ref48","doi-asserted-by":"publisher","first-page":"2278","DOI":"10.1109\/5.726791","article-title":"Gradient-based learning applied to document recognition","volume":"86","author":"LeCun","year":"1998","journal-title":"Proc of the IEEE"},{"key":"2025010523430736800_ref49","first-page":"1","article-title":"Learning multiple layers of features from tiny images","volume":"1","author":"Krizhevsky","year":"2009","journal-title":"Handb Syst Autoimm Dis"},{"key":"2025010523430736800_ref50","first-page":"1106","article-title":"Imagenet classification with deep convolutional neural networks","volume-title":"Proc. 26th Annu. Conf. Neural Inf. Process. Syst., Lake Tahoe, Nevada, United States, 3\u20136 December","author":"Krizhevsky","year":"2012"},{"key":"2025010523430736800_ref51","first-page":"429","article-title":"Federated optimization in heterogeneous networks","volume-title":"Proc. 3rd Conf. Mach. Learn. Syst., Austin, TX, USA, 2\u20134 March","author":"Li","year":"2020"},{"key":"2025010523430736800_ref52","first-page":"2021","article-title":"Fedpaq: a communication-efficient federated learning method with periodic averaging and quantization","volume-title":"Proc. 23rd Int. Conf. Artif. Intell. Stat., Palermo, Sicily, Italy, 26\u201328 August","author":"Reisizadeh","year":"2020"}],"container-title":["The Computer Journal"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/comjnl\/article-pdf\/67\/12\/3180\/59214800\/bxae081.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/comjnl\/article-pdf\/67\/12\/3180\/59214800\/bxae081.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,1,6]],"date-time":"2025-01-06T04:32:09Z","timestamp":1736137929000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/comjnl\/article\/67\/12\/3180\/7762888"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,9,20]]},"references-count":52,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2024,9,20]]},"published-print":{"date-parts":[[2024,12,20]]}},"URL":"https:\/\/doi.org\/10.1093\/comjnl\/bxae081","relation":{},"ISSN":["0010-4620","1460-2067"],"issn-type":[{"value":"0010-4620","type":"print"},{"value":"1460-2067","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2024,12]]},"published":{"date-parts":[[2024,9,20]]}}}