Name
Abstract | ||
|---|---|---|
The Electric Internet of Things (IoT) scenario generates a large amount of network edge data. The analysis of these data will significantly enhance the efficiency of IoT. However, traditional mobile edge computing (MEC) can not meet the need for data privacy protection. Currently, a new distributed learning approach, federated learning, has been proposed to solve data security problems and provide efficient model performance through the training of a shared model by multiple devices. Nevertheless, limited wireless communication resources and device heterogeneity are also key factors affecting FL performance Therefore, we propose a novel semi asynchronous federated learning (SAFL) framework, which aims to reduce calculation and transmission latency and improve the learning rate. We formulate a joint terminal device selection and resource allocation problem under the proposed SAFL framework to ensure communication efficiency. To deal with our proposed Markov Decision Process (MDP) problem, we introduce a deep deterministic policy gradient (DDPG) algorithm to find the optimal solution. Numerical results demonstrate the advantages of the proposed algorithm. |
Year | DOI | Venue |
|---|---|---|
2022 | 10.1109/CSCloud-EdgeCom54986.2022.00034 | 2022 IEEE 9th International Conference on Cyber Security and Cloud Computing (CSCloud)/2022 IEEE 8th International Conference on Edge Computing and Scalable Cloud (EdgeCom) |
Keywords | DocType | ISSN |
Federated learning,IoT,device selection,resource allocation,asynchronous | Conference | 2693-8952 |
ISBN | Citations | PageRank |
978-1-6654-8067-3 | 0 | 0.34 |
References | Authors | |
12 | 5 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Xian Luo | 1 | 0 | 0.68 |
| Zehua You | 2 | 0 | 0.34 |
| Rongtao Liao | 3 | 0 | 0.34 |
| Fen Liu | 4 | 0 | 0.34 |
| Liang Dong | 5 | 326 | 52.32 |