EUVIMEDEUVIMED entdecken
Crossref

Privacy-Preserving Federated Learning Based on Dataset Condensation

Di Zhang, Junqing Le, Nankun Mu, Gao Liu, Tao Xiang

IEEE Transactions on Consumer Electronics · 2025-02 · 71 (1) · 748-760

DOI: 10.1109/tce.2024.3510694 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Consumer Electronics, 2025-02. Verfasst von Di Zhang und 4 weiteren Autor:innen. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 53 Literaturangaben. Gefördert von: National Key Research and Development Program of China, Key Laboratory of Data Protection and Intelligent Management, Ministry of Education, Sichuan University, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Federated Learning (FL) is a decentralized learning framework that facilitates learning from large-scale datasets while keeping raw data locally. It dismantles data silos and addresses privacy concerns tied to direct data access. However, the substantial interaction of model parameters between local nodes and servers in FL significantly impacts training efficiency. Moreover, inversion attacks on model weights may also lead to serious privacy leakage. Towards this end, we combine data condensation and differential privacy (DP) techniques, designing a communication-efficient and privacy-preserving FL scheme named PPFL-DC. Specifically, the designed DP-processing method protects the model weights trained on the real dataset. Then, the protected model weights are used to guide the generation of lightweight and privacy-preserving synthetic images, where an efficient synthetic data generation method with an appropriate number of iterations is designed to reduce local computation costs. Instead of large-sized model weights, these small synthetic images are sent to the server. Subsequently, the server recovers the model weights from the synthetic images and uses them for model aggregation to achieve highly accurate global model updates. PPFL-DC effectively reduces communication and computation overhead, ensures client-level privacy protection, and maintains a high global model accuracy, even when dealing with Non-IID datasets. Finally, the theoretical analysis and experimental results on three benchmark datasets demonstrate the superiority of PPFL-DC in performance over the previous FL schemes.

Förderung

  • National Key Research and Development Program of China
  • Key Laboratory of Data Protection and Intelligent Management, Ministry of Education, Sichuan University
  • National Natural Science Foundation of China
  • Natural Science Foundation of Chongqing, China

Zitieren

APA

Zhang, D., Le, J., Mu, N., Liu, G., & Xiang, T. (2025). Privacy-Preserving Federated Learning Based on Dataset Condensation. IEEE Transactions on Consumer Electronics, 71(1), 748-760. https://doi.org/10.1109/tce.2024.3510694

Vancouver

Zhang D, Le J, Mu N, Liu G, Xiang T. Privacy-Preserving Federated Learning Based on Dataset Condensation. IEEE Transactions on Consumer Electronics. 2025;71(1):748-760. doi:10.1109/tce.2024.3510694

BibTeX
@article{Zhang2025,
  author = {Zhang, Di and Le, Junqing and Mu, Nankun and Liu, Gao and Xiang, Tao},
  title = {Privacy-Preserving Federated Learning Based on Dataset Condensation},
  journal = {IEEE Transactions on Consumer Electronics},
  year = {2025},
  volume = {71},
  number = {1},
  pages = {748--760},
  doi = {10.1109/tce.2024.3510694}
}
Kostenlos starten →