arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.00752cs.DC
收藏

区块链辅助的无线联邦学习:资源分配与客户端调度

Blockchain-aided wireless federated learning: Resource allocation and client scheduling

Jun Li, Weiwei Zhang, Kang Wei, Guangji Chen, Feng Shu, Wen Chen, Shi Jin

AI总结:

针对无线网络中区块链辅助去中心化联邦学习面临的资源受限与训练延迟问题,提出DRC-BDFL算法联合优化资源分配与客户端调度,在保证准确率相当的同时显著降低了训练延迟。

摘要:

基于中心化设计的联邦学习(FL)面临信任问题和单点故障的挑战。为了缓解这些问题,区块链辅助的去中心化联邦学习(BDFL)将去中心化网络架构引入FL训练过程,能有效克服中心化架构的缺陷。然而,在无线网络中部署BDFL通常会遇到带宽、计算能力和能耗受限等挑战。受此驱动,本文构建了一个动态随机优化问题,旨在通过联合优化资源分配和客户端选择,在有限能量预算和客户端参与的约束下,最小化平均训练延迟。我们利用Lyapunov优化工具求解该长期混合整数非线性规划问题,进而提出了动态资源分配与客户端调度的BDFL(DRC-BDFL)算法。此外,我们分析了DRC-BDFL的学习性能,并推导了关于全局损失函数的收敛上界。在SVHN和CIFAR-10数据集上进行的大量实验表明,DRC-BDFL在实现与基线算法相当准确率的同时,分别将训练延迟显著降低了9.24%和12.47%。

英文摘要:

Federated learning (FL) based on the centralized design faces both challenges regarding the trust issue and a single point of failure. To alleviate these issues, blockchain-aided decentralized FL (BDFL) introduces the decentralized network architecture into the FL training process, which can effectively overcome the defects of centralized architecture. However, deploying BDFL in wireless networks usually encounters challenges such as limited bandwidth, computing power, and energy consumption. Driven by these considerations, a dynamic stochastic optimization problem is formulated to minimize the average training delay by jointly optimizing the resource allocation and client selection under the constraints of limited energy budget and client participation. We solve the long-term mixed integer non-linear programming problem by employing the tool of Lyapunov optimization and thereby propose the dynamic resource allocation and client scheduling BDFL (DRC-BDFL) algorithm. Furthermore, we analyze the learning performance of DRC-BDFL and derive an upper bound for convergence regarding the global loss function. Extensive experiments conducted on SVHN and CIFAR-10 datasets demonstrate that DRC-BDFL achieves comparable accuracy to baseline algorithms while significantly reducing the training delay by 9.24% and 12.47%, respectively.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑