Dynamic user clustering for efficient and privacy-preserving federated learning
With the wider adoption of machine learning and increasing concern about data privacy, federated learning (FL) has received tremendous attention. FL schemes typically enable a set of participants, i.e., data owners, to individually train a machine learning model using their local data, which are the...
| Main Authors: | , , , , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/179908 |
| _version_ | 1811681930969612288 |
|---|---|
| author | Liu, Ziyao Guo, Jiale Yang, Wenzhuo Fan, Jiani Lam, Kwok-Yan Zhao, Jun |
| author2 | College of Computing and Data Science |
| author_facet | College of Computing and Data Science Liu, Ziyao Guo, Jiale Yang, Wenzhuo Fan, Jiani Lam, Kwok-Yan Zhao, Jun |
| author_sort | Liu, Ziyao |
| collection | NTU |
| description | With the wider adoption of machine learning and increasing concern about data privacy, federated learning (FL) has received tremendous attention. FL schemes typically enable a set of participants, i.e., data owners, to individually train a machine learning model using their local data, which are then aggregated with the coordination of a central server to construct a global FL model. Improvements upon standard FL include (i) reducing the communication overheads of gradient transmission by utilizing gradient sparsification and (ii) enhancing the security of aggregation by adopting privacy-preserving aggregation (PPAgg) protocols. However, state-of-the-art PPAgg protocols do not interoperate easily with gradient sparsification due to the heterogeneity of users' sparsified gradient vectors. To resolve this issue, we propose a Dynamic User Clustering (DUC) approach with a set of supporting protocols to partition users into clusters based on the nature of the PPAgg protocol and gradient sparsification technique, providing both security guarantees and communication efficiency. Experimental results show that DUC-FL significantly reduces communication overheads and achieves similar model accuracy compared to the baselines. The simplicity of the proposed protocol makes it attractive for both implementation and further improvements. |
| first_indexed | 2024-10-01T03:48:47Z |
| format | Journal Article |
| id | ntu-10356/179908 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T03:48:47Z |
| publishDate | 2024 |
| record_format | dspace |
| spelling | ntu-10356/1799082024-09-03T00:36:18Z Dynamic user clustering for efficient and privacy-preserving federated learning Liu, Ziyao Guo, Jiale Yang, Wenzhuo Fan, Jiani Lam, Kwok-Yan Zhao, Jun College of Computing and Data Science School of Computer Science and Engineering Digital Trust Centre Strategic Centre for Research in Privacy-Preserving Technologies & Systems (SCRIPTS) Computer and Information Science Federated learning Privacy preserving aggregation Gradient sparsification User clustering With the wider adoption of machine learning and increasing concern about data privacy, federated learning (FL) has received tremendous attention. FL schemes typically enable a set of participants, i.e., data owners, to individually train a machine learning model using their local data, which are then aggregated with the coordination of a central server to construct a global FL model. Improvements upon standard FL include (i) reducing the communication overheads of gradient transmission by utilizing gradient sparsification and (ii) enhancing the security of aggregation by adopting privacy-preserving aggregation (PPAgg) protocols. However, state-of-the-art PPAgg protocols do not interoperate easily with gradient sparsification due to the heterogeneity of users' sparsified gradient vectors. To resolve this issue, we propose a Dynamic User Clustering (DUC) approach with a set of supporting protocols to partition users into clusters based on the nature of the PPAgg protocol and gradient sparsification technique, providing both security guarantees and communication efficiency. Experimental results show that DUC-FL significantly reduces communication overheads and achieves similar model accuracy compared to the baselines. The simplicity of the proposed protocol makes it attractive for both implementation and further improvements. Info-communications Media Development Authority (IMDA) National Research Foundation (NRF) Submitted/Accepted version This research is supported by the National Research Foundation, Singapore and Infocomm Media Development Authority under its Trust Tech Funding Initiative and Strategic Capability Research Centres Funding Initiative. 2024-09-03T00:36:18Z 2024-09-03T00:36:18Z 2024 Journal Article Liu, Z., Guo, J., Yang, W., Fan, J., Lam, K. & Zhao, J. (2024). Dynamic user clustering for efficient and privacy-preserving federated learning. IEEE Transactions On Dependable and Secure Computing. https://dx.doi.org/10.1109/TDSC.2024.3355458 1545-5971 https://hdl.handle.net/10356/179908 10.1109/TDSC.2024.3355458 2-s2.0-85182922973 en IEEE Transactions on Dependable and Secure Computing © 2024 IEEE. All rights reserved. This article may be downloaded for personal use only. Any other use requires prior permission of the copyright holder. The Version of Record is available online at http://doi.org/10.1109/TDSC.2024.3355458. application/pdf |
| spellingShingle | Computer and Information Science Federated learning Privacy preserving aggregation Gradient sparsification User clustering Liu, Ziyao Guo, Jiale Yang, Wenzhuo Fan, Jiani Lam, Kwok-Yan Zhao, Jun Dynamic user clustering for efficient and privacy-preserving federated learning |
| title | Dynamic user clustering for efficient and privacy-preserving federated learning |
| title_full | Dynamic user clustering for efficient and privacy-preserving federated learning |
| title_fullStr | Dynamic user clustering for efficient and privacy-preserving federated learning |
| title_full_unstemmed | Dynamic user clustering for efficient and privacy-preserving federated learning |
| title_short | Dynamic user clustering for efficient and privacy-preserving federated learning |
| title_sort | dynamic user clustering for efficient and privacy preserving federated learning |
| topic | Computer and Information Science Federated learning Privacy preserving aggregation Gradient sparsification User clustering |
| url | https://hdl.handle.net/10356/179908 |
| work_keys_str_mv | AT liuziyao dynamicuserclusteringforefficientandprivacypreservingfederatedlearning AT guojiale dynamicuserclusteringforefficientandprivacypreservingfederatedlearning AT yangwenzhuo dynamicuserclusteringforefficientandprivacypreservingfederatedlearning AT fanjiani dynamicuserclusteringforefficientandprivacypreservingfederatedlearning AT lamkwokyan dynamicuserclusteringforefficientandprivacypreservingfederatedlearning AT zhaojun dynamicuserclusteringforefficientandprivacypreservingfederatedlearning |