EEFL: High-Speed Wireless Communications Inspired Energy Efficient Federated Learning over Mobile Devices
EEFL: High-Speed Wireless Communications Inspired Energy Efficient Federated Learning over Mobile Devices
复制标题
DOI：
10.1145/3581791.3596865
复制
发表时间：
2023-06
期刊：
影响因子：
--
通讯作者：
Rui Chen;Qiyu Wan;Xinyue Zhang;Xiaoqi Qin;Yanzhao Hou;D. Wang;Xin Fu;Miao Pan
中科院分区：
文献类型：
--
作者：
Rui Chen;Qiyu Wan;Xinyue Zhang;Xiaoqi Qin;Yanzhao Hou;D. Wang;Xin Fu;Miao Pan
能量效率对于移动的设备上的联邦学习及其潜在的繁荣应用至关重要。与将通信能耗视为瓶颈的现有通信高效FL研究工作不同，我们已经观察到，随着无线传输速度的不断增加（例如，Wi-Fi 5或5G），FL中用于模型更新的无线通信的能耗显著降低，有时甚至小于本地设备上训练的能耗。基于这种观察，在本文中，我们提出了一种基于高速无线通信的移动的设备节能联邦学习（EEFL），其目标是降低整体能耗（计算+通信）。特别是，我们设计了一种新的能量感知的自适应本地更新策略的移动的设备，通过联合考虑FL性能和节能的高速无线传输。此外，考虑到设备在每个FL全局轮中的本地更新策略，我们提出了动态电压和频率缩放（DVFS）策略，通过保持GPU和CPU以适当的频率工作而不触发热节流来最大限度地减少本地训练的能耗。各种学习模型、数据集和无线传输环境的广泛实验结果表明，所提出的EEFL在能源效率方面优于同类设计。
Energy efficiency is essential for federated learning (FL) over mobile devices and its potential prosperous applications. Different from existing communication efficient FL research efforts, which regard communication energy consumption as the bottleneck, we have observed that with ever increasing wireless transmission speed (e.g., Wi-Fi 5 or 5G), the energy consumption of wireless communications for model updates in FL is significantly reduced and sometimes is smaller than that of local on-device training. Motivated by such observations, in this paper, we propose a high-speed wireless communications inspired energy efficient federated learning over mobile devices (EEFL), whose goal is to reduce the overall energy consumption (computing + communication). In particular, we design a novel energy-aware adaptive local update policy for mobile devices by jointly considering FL performance and energy saving of high-speed wireless transmissions. Furthermore, given the device's local update policy in each FL global round, we advance the dynamic voltage and frequency scaling (DVFS) strategy to minimize local training's energy consumption by keeping GPU and CPU working at appropriate frequencies without triggering thermal throttling. Extensive experimental results with various learning models, datasets, and wireless transmission environments demonstrate the proposed EEFL's superiority over the peer designs in terms of energy efficiency.