EUVIMEDEUVIMED entdecken
Crossref

Efficient Federated Learning With Enhanced Privacy via Lottery Ticket Pruning in Edge Computing

Yifan Shi, Kang Wei, Li Shen, Jun Li, Xueqian Wang, Bo Yuan, Song Guo

IEEE Transactions on Mobile Computing · 2024-10 · 23 (10) · 9946-9958

DOI: 10.1109/tmc.2024.3370967 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Mobile Computing, 2024-10. Verfasst von Yifan Shi und 6 weiteren Autor:innen. Die Arbeit wurde bisher 26-mal zitiert (Crossref) und nennt 48 Literaturangaben. Gefördert von: National Key R&D Program of China, National Natural Science Foundation of China, Key Technologies R&D Program of Jiangsu.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Federated learning (FL) can train collaboratively with several mobile terminals (MTs), which faces critical challenges in communication, resource, and privacy. Existing privacy-preserving methods usually adopt instance-level differential privacy (DP), which provides a rigorous privacy guarantee but with several bottlenecks: performance degradation, transmission overhead, and resource constraints. Therefore, we propose Fed-LTP, an efficient and privacy-enhanced FL framework with Lottery Ticket Hypothesis (LTH) and zero-concentrated DP (zCDP). It generates a pruned global model on the server side and conducts sparse-to-sparse training from scratch with zCDP on the client side. On the server side, two pruning schemes are proposed: (i) the weight-based pruning (LTH) determines the pruned global model structure; (ii) the iterative pruning further shrinks the size of the pruned model. Meanwhile, the performance of Fed-LTP is boosted via model validation based on the Laplace mechanism. On the client side, we use sparse-to-sparse training to solve the resource-constraints issue and provide tighter privacy analysis to reduce the privacy budget. We evaluate the effectiveness of Fed-LTP on several real-world datasets in both independent and identically distributed (IID) and non-IID settings. The results confirm the superiority of Fed-LTP over state-of-the-art (SOTA) methods in communication, computation, and memory efficiencies while realizing a better utility-privacy trade-off.

Förderung

  • National Key R&D Program of China
  • National Natural Science Foundation of China
  • Key Technologies R&D Program of Jiangsu
  • Key-Area Research and Development Program of Guangdong Province
  • Hong Kong RGC Research Impact Fund
  • Areas of Excellence Scheme
  • General Research Fund
  • Shenzhen Science and Technology Innovation Commission

Zitieren

APA

Shi, Y., Wei, K., Shen, L., Li, J., Wang, X., Yuan, B., & Guo, S. (2024). Efficient Federated Learning With Enhanced Privacy via Lottery Ticket Pruning in Edge Computing. IEEE Transactions on Mobile Computing, 23(10), 9946-9958. https://doi.org/10.1109/tmc.2024.3370967

Vancouver

Shi Y, Wei K, Shen L, Li J, Wang X, Yuan B, et al. Efficient Federated Learning With Enhanced Privacy via Lottery Ticket Pruning in Edge Computing. IEEE Transactions on Mobile Computing. 2024;23(10):9946-9958. doi:10.1109/tmc.2024.3370967

BibTeX
@article{Shi2024,
  author = {Shi, Yifan and Wei, Kang and Shen, Li and Li, Jun and Wang, Xueqian and Yuan, Bo and Guo, Song},
  title = {Efficient Federated Learning With Enhanced Privacy via Lottery Ticket Pruning in Edge Computing},
  journal = {IEEE Transactions on Mobile Computing},
  year = {2024},
  volume = {23},
  number = {10},
  pages = {9946--9958},
  doi = {10.1109/tmc.2024.3370967}
}
Kostenlos starten →