{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T14:37:00Z","timestamp":1781534220619,"version":"3.54.5"},"reference-count":58,"publisher":"MDPI AG","issue":"5","license":[{"start":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T00:00:00Z","timestamp":1777593600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62302540"],"award-info":[{"award-number":["62302540"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"award":["62302540"],"award-info":[{"award-number":["62302540"]}],"id":[{"id":"https:\/\/ror.org\/01h0zpd94","id-type":"ROR","asserted-by":"publisher"}]},{"name":"the Key Research and Development Program of Henan Province","award":["251111212000"],"award-info":[{"award-number":["251111212000"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Computers"],"abstract":"<jats:p>Federated learning, a decentralized machine learning framework, allows multiple participants to jointly train models while keeping their raw data local and unshared. Nevertheless, during the exchange of model updates, the communicated information can still introduce privacy vulnerabilities and potentially result in the exposure of user data. Over the past few years, differential privacy methods have been broadly incorporated into federated learning frameworks to strengthen the protection of sensitive data. Nevertheless, the noise required to satisfy differential privacy guarantees often causes significant degradation in model performance. Prior studies have typically employed a fixed noise-injection strategy following gradient clipping. Although such methods provide privacy protection, they overlook the varying importance of different gradient dimensions, resulting in noise being injected into unimportant or redundant parameters, thereby causing unnecessary performance loss. To address these limitations, we propose an adaptive differentially private federated learning scheme with importance-based gradient updates (AIGU-DPFL). Specifically, we focus on coordinates with high information content and introduce an adaptive noise injection mechanism, which perturbs gradient updates to satisfy differential privacy guarantees while dynamically controlling noise intensity, thereby achieving sparse and noise-effective gradient updates. AIGU-DPFL markedly enhances the training effectiveness of federated learning models. Comprehensive evaluations conducted on real-world datasets indicate that the proposed method achieves superior performance compared to existing differentially private federated learning techniques.<\/jats:p>","DOI":"10.3390\/computers15050288","type":"journal-article","created":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T16:54:07Z","timestamp":1777654447000},"page":"288","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["AIGU-DPFL: Adaptive Differentially Private Federated Learning with Importance-Based Gradient Updates"],"prefix":"10.3390","volume":"15","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5586-6715","authenticated-orcid":false,"given":"Fangfang","family":"Shan","sequence":"first","affiliation":[{"name":"College of Computer, Zhongyuan University of Technology, Zhengzhou 450007, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-3812-6157","authenticated-orcid":false,"given":"Zhuo","family":"Chen","sequence":"additional","affiliation":[{"name":"College of Computer, Zhongyuan University of Technology, Zhengzhou 450007, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yifan","family":"Mao","sequence":"additional","affiliation":[{"name":"College of Computer, Zhongyuan University of Technology, Zhengzhou 450007, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-8950-2668","authenticated-orcid":false,"given":"Yuhang","family":"Liu","sequence":"additional","affiliation":[{"name":"College of Computer, Zhongyuan University of Technology, Zhengzhou 450007, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lulu","family":"Fan","sequence":"additional","affiliation":[{"name":"College of Computer, Zhongyuan University of Technology, Zhengzhou 450007, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yanlong","family":"Lu","sequence":"additional","affiliation":[{"name":"College of Computer, Zhongyuan University of Technology, Zhengzhou 450007, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,5,1]]},"reference":[{"key":"ref_1","first-page":"1273","article-title":"Communication-efficient learning of deep networks from decentralized data","volume":"Volume 54","author":"McMahan","year":"2017","journal-title":"Proceedings of the 20th International Conference on Artificial Intelligence and Statistics"},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Long, G., Tan, Y., and Jiang, J. (2020). Federated learning for open banking. Federated Learning: Privacy and Incentive, Springer.","DOI":"10.1007\/978-3-030-63076-8_17"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"535","DOI":"10.1109\/ACCESS.2021.3137364","article-title":"Reliable and resilient AI and IoT-based personalized healthcare services: A survey","volume":"10","author":"Taimoor","year":"2021","journal-title":"IEEE Access"},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"2648","DOI":"10.1109\/TIFS.2016.2594132","article-title":"Reconstruction attacks against mobile-based continuous authentication systems in the cloud","volume":"11","author":"Chang","year":"2016","journal-title":"IEEE Trans. Inf. Forensics Secur."},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Shokri, R., Stronati, M., Song, C., and Shmatikov, V. (2017, January 22\u201326). Membership inference attacks against machine learning models. Proceedings of the 2017 IEEE Symposium on Security and Privacy (SP), San Jose, CA, USA.","DOI":"10.1109\/SP.2017.41"},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Salem, A., Zhang, Y., Humbert, M., Berrang, P., and Fritz, M. (2018). ML-Leaks: Model and data independent membership inference attacks and defenses. arXiv.","DOI":"10.14722\/ndss.2019.23119"},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Melis, L., Song, C., De Cristofaro, E., and Shmatikov, V. (2019, January 19\u201323). Exploiting unintended feature leakage in collaborative learning. Proceedings of the 2019 IEEE Symposium on Security and Privacy (SP), San Francisco, CA, USA.","DOI":"10.1109\/SP.2019.00029"},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Dwork, C. (2006). Differential privacy. International Colloquium on Automata, Languages, and Programming, Springer.","DOI":"10.1007\/11787006_1"},{"key":"ref_9","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":"ref_10","unstructured":"Wang, Z., Yu, X., Huang, Q., and Gong, Y. (2024). An adaptive differential privacy method based on federated learning. arXiv."},{"key":"ref_11","unstructured":"Geyer, R.C., Klein, T., and Nabi, M. (2017). Differentially private federated learning: A client level perspective. arXiv."},{"key":"ref_12","unstructured":"Choudhury, O., Gkoulalas-Divanis, A., Salonidis, T., Sylla, S., and Das, G. (2019). Differential privacy-enabled federated learning for sensitive health data. arXiv."},{"key":"ref_13","unstructured":"Zheng, Q., Chen, S., Long, Q., and Su, W. (2021, January 13\u201315). Federated f-differential privacy. Proceedings of the International Conference on Artificial Intelligence and Statistics, Virtual."},{"key":"ref_14","unstructured":"McMahan, H.B., Ramage, D., Talwar, K., and Zhang, L. (2017). Learning differentially private recurrent language models. arXiv."},{"key":"ref_15","doi-asserted-by":"crossref","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":"ref_16","doi-asserted-by":"crossref","unstructured":"Shokri, R., and Shmatikov, V. (2015, January 12\u201316). Privacy-preserving deep learning. Proceedings of the 22nd ACM SIGSAC Conference on Computer and Communications Security, Denver, CO, USA.","DOI":"10.1145\/2810103.2813687"},{"key":"ref_17","unstructured":"Jayaraman, B., and Evans, D. (2019, January 14\u201316). Evaluating differentially private machine learning in practice. Proceedings of the 28th USENIX Security Symposium, Santa Clara, CA, USA."},{"key":"ref_18","unstructured":"Xiang, L., Yang, J., and Li, B. (May, January 29). Differentially-private deep learning from an optimization perspective. Proceedings of the IEEE INFOCOM, Paris, France."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"8242","DOI":"10.1109\/TMC.2023.3343288","article-title":"Federated learning with sparsified model perturbation: Improving accuracy under client-level differential privacy","volume":"23","author":"Hu","year":"2023","journal-title":"IEEE Trans. Mob. Comput."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"14722","DOI":"10.1109\/TMC.2024.3445957","article-title":"P2CEFL: Privacy-preserving and communication efficient federated learning with sparse gradient and dithering quantization","volume":"23","author":"Wang","year":"2024","journal-title":"IEEE Trans. Mob. Comput."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"142303","DOI":"10.1007\/s11432-023-3918-9","article-title":"Gradient sparsification for efficient wireless federated learning with differential privacy","volume":"67","author":"Wei","year":"2024","journal-title":"Sci. China Inf. Sci."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Medjadji, C., Alawadi, S., Awaysheh, F.M., Leduc, G., Kubler, S., and Le Traon, Y. (2025, January 10\u201312). FedSparQ: Adaptive sparse quantization with error feedback for robust and efficient federated learning. Proceedings of the 3rd International Conference on Federated Learning Technologies and Applications (FLTA), Valencia, Spain.","DOI":"10.1109\/FLTA67013.2025.11336403"},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"1109","DOI":"10.1007\/s11227-025-07593-0","article-title":"AdaS-FLDP: Local differentially private federated learning with adaptive sparsification","volume":"81","author":"Huang","year":"2025","journal-title":"J. Supercomput."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"1243","DOI":"10.1109\/TPAMI.2023.3332428","article-title":"Multi-stage asynchronous federated learning with adaptive differential privacy","volume":"46","author":"Li","year":"2023","journal-title":"IEEE Trans. Pattern. Anal. Mach. Intell."},{"key":"ref_25","unstructured":"Andrew, G., Thakkar, O., McMahan, B., and Ramaswamy, S. (2021, January 6\u201314). Differentially private learning with adaptive clipping. Proceedings of the Advances in Neural Information Processing Systems, Virtual."},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Fu, J., Chen, Z., and Han, X. (2022, January 17\u201319). Adap DP-FL: Differentially private federated learning with adaptive noise. Proceedings of the 2022 IEEE International Conference on Trust, Security and Privacy in Computing and Communications, Wuhan, China.","DOI":"10.1109\/TrustCom56396.2022.00094"},{"key":"ref_27","first-page":"142303","article-title":"Federated learning design and functional models: Survey","volume":"67","author":"Ayeelyan","year":"2025","journal-title":"Artif. Intell. Rev."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"153817","DOI":"10.1109\/ACCESS.2025.3605165","article-title":"Advancing federated learning: A systematic literature review of methods, challenges, and applications","volume":"13","author":"Sana","year":"2025","journal-title":"IEEE Access"},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Xu, Z., Shi, S., Liu, A.X., Zhao, J., and Chen, L. (2020, January 6\u20139). An adaptive and fast convergent approach to differentially private deep learning. Proceedings of the IEEE INFOCOM 2020, Virtual.","DOI":"10.1109\/INFOCOM41043.2020.9155359"},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Yu, L., Liu, L., Pu, C., Gursoy, M.E., and Truex, S. (2019, January 19\u201323). Differentially private model publishing for deep learning. Proceedings of the 2019 IEEE Symposium on Security and Privacy (SP), San Francisco, CA, USA.","DOI":"10.1109\/SP.2019.00019"},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Lee, J., and Kifer, D. (2018, January 19\u201323). Concentrated differentially private gradient descent with adaptive per-iteration privacy budget. Proceedings of the 24th ACM SIGKDD Conference, London, UK.","DOI":"10.1145\/3219819.3220076"},{"key":"ref_32","unstructured":"Pichapati, V., Suresh, A.T., Yu, F.X., Reddi, S.J., and Kumar, S. (2019). AdaClip: Adaptive clipping for private SGD. arXiv."},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"74","DOI":"10.1109\/TIFS.2023.3318944","article-title":"Differentially private federated learning with an adaptive noise mechanism","volume":"19","author":"Xue","year":"2023","journal-title":"IEEE Trans. Inf. Forensics Secur."},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"303","DOI":"10.1109\/TIFS.2021.3139777","article-title":"Gradient leakage attack resilient deep learning","volume":"17","author":"Wei","year":"2021","journal-title":"IEEE Trans. Inf. Forensics Secur."},{"key":"ref_35","unstructured":"Tramer, F., and Boneh, D. (2020). Differentially private learning needs better features (or much more data). arXiv."},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"2208","DOI":"10.1109\/TPAMI.2018.2855738","article-title":"Scattering networks for hybrid representation learning","volume":"41","author":"Oyallon","year":"2018","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_37","unstructured":"Sha, H., Liu, R., Liu, Y., and Chen, H. (2023). PCDP-SGD: Improving the convergence of differentially private SGD via projection in advance. arXiv."},{"key":"ref_38","unstructured":"Liu, Y., Xiong, L., Liu, Y., Gu, Y., Liu, R., and Chen, H. (2024). DPDR: Gradient decomposition and reconstruction for differentially private deep learning. arXiv."},{"key":"ref_39","unstructured":"Zhu, J., and Blaschko, M.B. (2021). Improving differentially private SGD via randomly sparsified gradients. arXiv."},{"key":"ref_40","unstructured":"Adamczewski, K., and Park, M. (2023). Differential privacy meets neural network pruning. arXiv."},{"key":"ref_41","doi-asserted-by":"crossref","first-page":"1770","DOI":"10.1109\/TKDE.2018.2805356","article-title":"Privacy enhanced matrix factorization for recommendation with local differential privacy","volume":"30","author":"Shin","year":"2018","journal-title":"IEEE Trans. Knowl. Data Eng."},{"key":"ref_42","unstructured":"Kirszbraun, M.D. (1934). Extensions of Lipschitz mappings into a Hilbert space. Ann. Math., 45\u201376."},{"key":"ref_43","doi-asserted-by":"crossref","unstructured":"Mironov, I. (2017, January 21\u201325). R\u00e9nyi differential privacy. Proceedings of the 2017 IEEE 30th Computer Security Foundations Symposium (CSF), Santa Barbara, CA, USA.","DOI":"10.1109\/CSF.2017.11"},{"key":"ref_44","doi-asserted-by":"crossref","unstructured":"Dwork, C., McSherry, F., Nissim, K., and Smith, A. (2006). Our data, ourselves: Privacy via distributed noise generation. Advances in Cryptology\u2014EUROCRYPT, Springer.","DOI":"10.1007\/11761679_29"},{"key":"ref_45","doi-asserted-by":"crossref","unstructured":"Dwork, C., McSherry, F., Nissim, K., and Smith, A. (2006). Calibrating noise to sensitivity in private data analysis. Theory of Cryptography Conference, Springer.","DOI":"10.1007\/11681878_14"},{"key":"ref_46","doi-asserted-by":"crossref","first-page":"3635","DOI":"10.1109\/TCE.2023.3338464","article-title":"Differentially private federated learning with importance client sampling","volume":"70","author":"Chen","year":"2023","journal-title":"IEEE Trans. Consum. Electron"},{"key":"ref_47","unstructured":"Zhang, J., He, T., Sra, S., and Jadbabaie, A. (2019). Why gradient clipping accelerates training: A theoretical justification for adaptivity. arXiv."},{"key":"ref_48","unstructured":"Loshchilov, I., and Hutter, F. (2016). SGDR: Stochastic gradient descent with warm restarts. arXiv."},{"key":"ref_49","unstructured":"Morales-Brotons, D., Vogels, T., and Hendrikx, H. (2024). Exponential moving average of weights in deep learning: Dynamics and benefits. arXiv."},{"key":"ref_50","doi-asserted-by":"crossref","first-page":"141","DOI":"10.1109\/MSP.2012.2211477","article-title":"The MNIST database of handwritten digit images for machine learning research","volume":"29","author":"Deng","year":"2012","journal-title":"IEEE Signal Process Mag."},{"key":"ref_51","unstructured":"Xiao, H., Rasul, K., and Vollgraf, R. (2017). Fashion-MNIST: A novel image dataset for benchmarking machine learning algorithms. arXiv."},{"key":"ref_52","unstructured":"Krizhevsky, A., and Hinton, G. (2009). Learning Multiple Layers of Features from Tiny Images, University of Toronto. Technical Report."},{"key":"ref_53","doi-asserted-by":"crossref","unstructured":"Qiu, C., Wu, Z., Wang, H., Yang, Q., Wang, Y., and Su, C. (2025). Hierarchical aggregation for federated learning in heterogeneous IoT scenarios: Enhancing privacy and communication efficiency. Future Internet, 17.","DOI":"10.3390\/fi17010018"},{"key":"ref_54","doi-asserted-by":"crossref","unstructured":"Liu, P., Jia, L., and Xiao, Y. (2025). Participant selection for efficient and trusted federated learning in blockchain-assisted hierarchical federated learning architectures. Future Internet, 17.","DOI":"10.3390\/fi17020075"},{"key":"ref_55","doi-asserted-by":"crossref","unstructured":"He, Z., Zhu, G., Zhang, S., Luo, E., and Zhao, Y. (2025). FedDT: A communication-efficient federated learning via knowledge distillation and ternary compression. Electronics, 14.","DOI":"10.3390\/electronics14112183"},{"key":"ref_56","unstructured":"Lin, Y., Han, S., Mao, H., Wang, Y., and Dally, W.J. (2017). Deep gradient compression: Reducing the communication bandwidth for distributed training. arXiv."},{"key":"ref_57","doi-asserted-by":"crossref","unstructured":"Zhang, X., Ding, J., Wu, M., Wong, S.T., Van Nguyen, H., and Pan, M. (2021, January 3\u20138). Adaptive privacy preserving deep learning algorithms for medical data. Proceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision, Waikoloa, HI, USA.","DOI":"10.1109\/WACV48630.2021.00121"},{"key":"ref_58","doi-asserted-by":"crossref","first-page":"101476","DOI":"10.1016\/j.elerap.2025.101476","article-title":"DP-SGD-global-adapt-V2-S: Triad improvements of privacy, accuracy and fairness via step decay noise multiplier and step decay upper clipping threshold","volume":"70","author":"Chilukoti","year":"2025","journal-title":"Electron. Commer. Res. Appl."}],"container-title":["Computers"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2073-431X\/15\/5\/288\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,19]],"date-time":"2026-05-19T04:22:42Z","timestamp":1779164562000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2073-431X\/15\/5\/288"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,1]]},"references-count":58,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2026,5]]}},"alternative-id":["computers15050288"],"URL":"https:\/\/doi.org\/10.3390\/computers15050288","relation":{},"ISSN":["2073-431X"],"issn-type":[{"value":"2073-431X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,1]]}}}