CoChat
Sign Up
Cited by 1,407Privacy-Preserving Technologies in DataCloud Data Security SolutionsStochastic Gradient Optimization Techniques

A Survey on Federated Learning Systems: Vision, Hype and Reality for Data Privacy and Protection

Open access
Li, Qinbin, Wen, Zeyi, Wu, Zhaomin, Hu, Sixu, Wang, Naibo, Li, Yuan, Liu, Xu, He, Bingsheng·2023·IEEE Transactions on Knowledge and Data Engineering
DOI 10.1109/tkde.2021.3124599
Open this paper in CoChat — free accountOpen PDF

Abstract

As data privacy increasingly becomes a critical societal concern, federated learning has been a hot research topic in enabling the collaborative training of machine learning models among different organizations under the privacy restrictions. As researchers try to support more machine learning models with different privacy-preserving approaches, there is a requirement in developing systems and infrastructures to ease the development of various federated learning algorithms. Similar to deep learning systems such as PyTorch and TensorFlow that boost the development of deep learning, federated learning systems (FLSs) are equivalently important, and face challenges from various aspects such as effectiveness, efficiency, and privacy. In this survey, we conduct a comprehensive review on federated learning systems. To understand the key design system components and guide future research, we introduce the definition of federated learning systems and analyze the system components. Moreover, we provide a thorough categorization for federated learning systems according to six different aspects, including data distribution, machine learning model, privacy mechanism, communication architecture, scale of federation and motivation of federation. The categorization can help the design of federated learning systems as shown in our case studies. By systematically summarizing the existing federated learning systems, we present the design factors, case studies, and future research opportunities.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/tkde.2021.3124599

Cites

216 of 216 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Differential privacy is applied on shared parameters with a Gaussian mechanism to provide strict privacy preservation on a proxy server as the middle layer between the server and all the participants.

    “Differential privacy is then applied on shared parameters with a Gaussian mechanism to provide strict privacy preservation on a proxy server as the middle layer between the server and all the participants.”

    From Federated Learning for Internet of Things: A Comprehensive Survey (2021) · cites this paper as [190]

    Section 4.2.3 (p. 16) describes a Gaussian mechanism applied to aggregated gradient updates for differential privacy, and Section 3.3 (p. 8) describes adding noise to model parameters for privacy protection. The cited passages do not describe a proxy server between the central server and participants or place that mechanism on one. (LLM, conf=0.86)

    Experimental — accuracy not yet measured

  • To mitigate privacy leakage, fewer parameters are shared between the server and each participant while having less impact on the model accuracy.

    “To mitigate privacy leakage, fewer parameters are shared between the server and each participant while having less impact on the model accuracy.”

    From Federated Learning for Internet of Things: A Comprehensive Survey (2021) · cites this paper as [190]

    Section 3.3 (p. 8) says exchanged model parameters may leak sensitive information. Section 4.2.2 (p. 16) describes sending a smaller update with only slight degradation in convergence speed, but presents this as a way to reduce communication costs—not as a demonstrated way to mitigate privacy leakage or preserve model accuracy. (LLM, conf=0.83)

    Experimental — accuracy not yet measured

  • An anonymous and privacy-preserving FL scheme is introduced for the mining of industrial big data.

    “Similarly, an anonymous and privacypreserving FL scheme is introduced in [190] for the mining of industrial big data.”

    From Federated Learning for Internet of Things: A Comprehensive Survey (2021) · cites this paper as [190]

    Section 4.3.1 (p. 21) describes FATE as an industrial-level FL framework that uses secure protocols to protect communicated messages. It does not describe FATE as anonymous or as a scheme for mining industrial big data, and identifies WeBank—not the survey authors—as its developer. (LLM, conf=0.82)

    Experimental — accuracy not yet measured

  • WFL cannot always guarantee the security of the training process.

    “Recent works have shown that WFL cannot always guarantee the security of the training process [46], [82], [93], [105], [106].”

    From Data and model poisoning backdoor attacks on wireless federated learning, and the defense mechanisms: A comprehensive survey (2024) · cites this paper as [93]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 19

  1. [18][18] Jeremy Bernstein, Yu-Xiang Wang, Kamyar Azizzadenesheli, and Anima Anandkumar. signsgd: Compressed optimisation for non-convex problems. arXiv preprint arXiv:1802.04434 , 2018. - · arXiv
  2. [23][23] Keith Bonawitz, Vladimir Ivanov, Ben Kreuter, Antonio Marcedone, H Brendan McMahan, Sarvar Patel, Daniel Ramage, Aaron Segal, and Karn Seth. Practical secure aggregation for privacypreserving machine learning. In Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security , pages 1175-1191. ACM, 2017. -
  3. [46][46] Canh T Dinh, Nguyen H Tran, and Tuan Dung Nguyen. Personalized federated learning with moreau envelopes. arXiv preprint arXiv:2006.08848 , 2020. - · arXiv
  4. [51][51] Ittay Eyal, Adem Efe Gencer, Emin Gun Sirer, and Robbert Van Renesse. Bitcoin-ng: A scalable blockchain protocol. In 13th USENIX Symposium on Networked Systems Design and Implementation (NSDI 16) , pages 45-59, Santa Clara, CA, March 2016. USENIX Association. ISBN 978-1931971-29-4. URL https://www.usenix.org/conference/nsdi16/technical-sessions/presentation/eyal. -
  5. [53][53] Minghong Fang, Xiaoyu Cao, Jinyuan Jia, and Neil Gong. Local model poisoning attacks to byzantine-robust federated learning. In USENIX , 2020. -
  6. [90][90] Sai Praneeth Karimireddy, Satyen Kale, Mehryar Mohri, Sashank Reddi, Sebastian Stich, and Ananda Theertha Suresh. Scaffold: Stochastic controlled averaging for federated learning. In International Conference on Machine Learning , pages 5132-5143. PMLR, 2020. -
  7. [119][119] Yang Liu, Tianjian Chen, and Qiang Yang. Secure federated transfer learning. arXiv preprint arXiv:1812.03337 , 2018. - · arXiv
  8. [123][123] Noel Lopes and Bernardete Ribeiro. Gpumlib: An efficient open-source gpu machine learning library. International Journal of Computer Information Systems and Industrial Management Applications , 3:355-362, 2011. -
  9. [128][128] Othmane Marfoq, Chuan Xu, Giovanni Neglia, and Richard Vidal. Throughput-optimal topology design for cross-silo federated learning. arXiv preprint arXiv:2010.12229 , 2020. - · arXiv
  10. [133][133] Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A Rusu, Joel Veness, Marc G Bellemare, Alex Graves, Martin Riedmiller, Andreas K Fidjeland, Georg Ostrovski, et al. Humanlevel control through deep reinforcement learning. Nature , 518(7540):529-533, 2015. -
  11. [156][156] MSadegh Riazi, Christian Weinert, Oleksandr Tkachenko, Ebrahim M Songhori, Thomas Schneider, and Farinaz Koushanfar. Chameleon: A hybrid secure computation framework for machine learning applications. In Proceedings of the 2018 on Asia Conference on Computer and Communications Security , pages 707-721. ACM, 2018. -
  12. [169][169] Virginia Smith, Chao-Kai Chiang, Maziar Sanjabi, and Ameet S Talwalkar. Federated multi-task learning. In Advances in Neural Information Processing Systems , pages 4424-4434, 2017. -
  13. [179][179] ADP Team et al. Learning with privacy at scale. Apple Machine Learning Journal , 1(8), 2017. -
  14. [182][182] Manasi Vartak, Harihar Subramanyam, Wei-En Lee, Srinidhi Viswanathan, Saadiyah Husnoo, Samuel Madden, and Matei Zaharia. Modeldb: a system for machine learning model management. In Proceedings of the Workshop on Human-In-the-Loop Data Analytics , pages 1-3, 2016. -
  15. [189][189] Hongyi Wang, Mikhail Yurochkin, Yuekai Sun, Dimitris Papailiopoulos, and Yasaman Khazaeni. Federated learning with matched averaging. arXiv preprint arXiv:2002.06440 , 2020. - · arXiv
  16. [190][190] Jianyu Wang, Qinghua Liu, Hao Liang, Gauri Joshi, and H Vincent Poor. Tackling the objective inconsistency problem in heterogeneous federated optimization. Advances in Neural Information Processing Systems , 2020. -
  17. [191][191] Rui Wang, Heju Li, and Erwu Liu. Blockchain-based federated learning in mobile edge networks with application in internet of vehicles. arXiv preprint arXiv:2103.01116 , 2021. - · arXiv
  18. [205][205] Runhua Xu, Nathalie Baracaldo, Yi Zhou, Ali Anwar, and Heiko Ludwig. Hybridalpha: An efficient approach for privacy-preserving federated learning. In Proceedings of the 12th ACM Workshop on Artificial Intelligence and Security , pages 13-23, 2019. -
  19. [213][213] Mikhail Yurochkin, Mayank Agarwal, Soumya Ghosh, Kristjan Greenewald, Trong Nghia Hoang, and Yasaman Khazaeni. Bayesian nonparametric federated learning of neural networks. arXiv preprint arXiv:1905.12022 , 2019. - · arXiv

Cited by

Sorted by citations

Related

Authors
Li, Q., Wen, Z., Wu, Z., Hu, S., Wang, N., Li, Y., Liu, X., & He, B.
Affiliations
  • National University of Singapore (SG)
  • The University of Western Australia (AU)
ORCID
Venue
IEEE Transactions on Knowledge and Data Engineering
Publisher
Institute of Electrical and Electronics Engineers (IEEE)
Published
2023-04-01
Volume
35
Issue
4
Pages
3347-3366
ISSN
  • 1041-4347 (ISSN-L)
  • 1558-2191
  • 2326-3865
DOI
10.1109/tkde.2021.3124599
OpenAlex
W2992272656
Cited by
1,407
References
216
Citations by year
2026: 239 · 2025: 382 · 2024: 370 · 2023: 227 · 2022: 117 · 2021: 55 · 2020: 16 · 2019: 1
FWCI
67.94
Type
journal article
Language
en
Open access
Licence
Locations
  • IEEE Transactions on Knowledge and Data Engineering (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • UWA Profiles and Research Repository (UWA) (submittedVersion)
  • Rare & Special e-Zone (The Hong Kong University of Science and Technology) (submittedVersion)
  • arXiv (Cornell University)
Funders
National Research Foundation Singapore (AISG2-RP-2020-018)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
  • Physical Sciences › Computer Science › Information Systems › Cloud Data Security Solutions
  • Physical Sciences › Computer Science › Artificial Intelligence › Stochastic Gradient Optimization Techniques
Keywords
data privacy, privacy-preserving machine learning, system architecture, data distribution, communication architecture, privacy mechanisms

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review