mlsanityData to 7 Oct 2026

Amplitude-Varying Perturbation for Balancing Privacy and Utility in Federated Learning

arXiv 2303.04274 ↗First posted 7 Mar 2023cs.LG cs.CR cs.PFcited by 13

Not on the map. Why?

Abstract (arXiv)

While preserving the privacy of federated learning (FL), differential privacy (DP) inevitably degrades the utility (i.e., accuracy) of FL due to model perturbations caused by DP noise added to model updates. Existing studies have considered exclusively noise with persistent root-mean-square amplitude and overlooked an opportunity of adjusting the amplitudes to alleviate the adverse effects of the noise. This paper presents a new DP perturbation mechanism with a time-varying noise amplitude to protect the privacy of FL and retain the capability of adjusting the learning performance. Specifically, we propose a geometric series form for the noise amplitude and reveal analytically the dependence of the series on the number of global aggregations and the (ε,δ)-DP requirement. We derive an online refinement of the series to prevent FL from premature convergence resulting from excessive perturbation noise. Another important aspect is an upper bound developed for the loss function of a multi-layer perceptron (MLP) trained by FL running the new DP mechanism. Accordingly, the optimal number of global aggregations is obtained, balancing the learning and privacy. Extensive experiments are conducted using MLP, supporting vector machine, and convolutional neural network models on four public datasets. The contribution of the new DP mechanism to the convergence and accuracy of privacy-preserving FL is corroborated, compared to the state-of-the-art Gaussian noise mechanism with a persistent noise amplitude.

Key facts. Cited by 13 papers in the map (data through 7 Oct 2026). Builds on 26 of its 42 references that are in the map. First posted on arXiv 7 Mar 2023.

Builds on

26 of 42 references.

References · 26

  1. 2016Deep Learning with Differential Privacycited by 2,682
  2. 2021Convex optimizationcited by 2,479
  3. 2016Xception: Deep Learning with Depthwise Separable Convolutionscited by 2,320
  4. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,221
  5. 2019Federated Learning: Challenges, Methods, and Future Directionscited by 1,531
  6. 2017Renyi Differential Privacycited by 643
  7. 2019Federated Learning with Differential Privacy: Algorithms and Performance Analysiscited by 528
  8. 2016Linear Convergence of Gradient and Proximal-Gradient Methods Under the Polyak-Lojasiewicz Conditioncited by 526
  9. 2017Differentially Private Federated Learning: A Client Level Perspectivecited by 451
  10. 2019Federated Machine Learning: Concept and Applicationscited by 409
  11. 2018Adaptive Federated Learning in Resource Constrained Edge Computing Systemscited by 408
  12. 2021Federated Learning with Buffered Asynchronous Aggregationcited by 188
14 more
  1. 2020LDP-Fed: Federated Learning with Local Differential Privacycited by 129
  2. 2018Protection Against Reconstruction and Its Applications in Private Federated Learningcited by 127
  3. 2014Differentially Private Empirical Risk Minimization: Efficient Algorithms and Tight Error Boundscited by 104
  4. 2021PPFL: Privacy-preserving Federated Learning with Trusted Execution Environmentscited by 93
  5. 2020Local Differential Privacy based Federated Learning for Internet of Thingscited by 74
  6. 2019Differentially Private Model Publishing for Deep Learningcited by 69
  7. 2019On Safeguarding Privacy and Security in the Framework of Federated Learningcited by 55
  8. 2017Differentially Private Mixture of Generative Neural Networkscited by 39
  9. 2022Differentially Private Federated Learning with Local Regularization and Sparsificationcited by 35
  10. 2021Low-Latency Federated Learning over Wireless Channels with Differential Privacycited by 24
  11. 2022How Much Privacy Does Federated Learning with Secure Aggregation Guarantee?cited by 19
  12. 2020Secure and Fault Tolerant Decentralized Learningcited by 6
  13. 2021On the Practicality of Differential Privacy in Federated Learning by Tuning Iteration Timescited by 2
  14. 2021Super-convergence and Differential Privacy: Training faster with better privacy guaranteescited by 1

Built on by

Cited by year

Citing papers per year: 2023: 4; 2024: 1; 2025: 8.

13 papers.

Sort:

By area:

  1. 2023Over-The-Air Federated Learning: Status Quo, Open Challenges, and Future Directionscited by 21
  2. 2023Data and Model Poisoning Backdoor Attacks on Wireless Federated Learning, and the Defense Mechanisms: A Comprehensive Surveycited by 17
  3. 2024Clients Collaborate: Flexible Differentially Private Federated Learning with Guaranteed Improvement of Utility-Privacy Trade-offcited by 6
  4. 2023Blockchained Federated Learning for Internet of Things: A Comprehensive Surveycited by 5
  5. 2025Responsible Diffusion: A Comprehensive Survey on Safety, Ethics, and Trust in Diffusion Modelscited by 3
  6. 2023Design of Two-Level Incentive Mechanisms for Hierarchical Federated Learningcited by 1
  7. 2025Zero-Trust Foundation Models: A New Paradigm for Secure and Collaborative Artificial Intelligence for Internet of Thingscited by 1
  8. 2025Enhancing Convergence, Privacy and Fairness for Wireless Personalized Federated Learning: Quantization-Assisted Min-Max Fair Schedulingcited by 1
  9. 2025Free Privacy Protection for Wireless Federated Learning: Enjoy It or Suffer from It?cited by 0
  10. 2025Convergence-Privacy-Fairness Trade-Off in Personalized Federated Learningcited by 0
  11. 2025Privacy-Preserving Quantized Federated Learning with Diverse Precisioncited by 0
  12. 2025ParaAegis: Parallel Protection for Flexible Privacy-preserved Federated Learningcited by 0
1 more
  1. 2025ADP-VRSGP: Decentralized Learning with Adaptive Differential Privacy via Variance-Reduced Stochastic Gradient Pushcited by 0