一种基于带宽分配的联邦学习激励机制

doi:10.11959/j.issn.2096-3750.2022.00300

物联网学报 ›› 2022, Vol. 6 ›› Issue (4): 82-92.doi: 10.11959/j.issn.2096-3750.2022.00300

一种基于带宽分配的联邦学习激励机制

郭英芸¹^,², 高博¹^,², 张志飞¹^,², 张煜³, 熊轲¹^,²

¹ 北京交通大学计算机与信息技术学院，北京 100044
² 北京交通大学高速铁路网络管理教育部工程研究中心，北京 100044
³ 国网能源研究院有限公司，北京 102209

修回日期:2022-09-16 出版日期:2022-12-30 发布日期:2022-12-01
作者简介:郭英芸（1995- ），女，北京交通大学硕士生，主要研究方向为移动与互联网络
高博（1984- ），男，北京交通大学副教授，主要研究方向为无线网络、移动计算、机器学习
张志飞（1971- ），男，博士，北京交通大学计算机与信息技术学院高级工程师，博士，主要研究方向为网络通信理论、网络安全等
张煜（1983- ），男，博士，国网能源研究院有限公司高级工程师，主要研究方向为泛在电力物联网、无线协作网络、电能替代等
熊轲（1981-），男，博士，北京交通大学计算机与信息技术学院教授、副院长，主要研究方向为无线协作网络、无线移动网络和网络信息理论等
基金资助:
国家自然科学基金资助项目(61872028);中央高校基本科研业务费专项资金(2021JBM008);中央高校基本科研业务费专项资金(2022JBXT001)

An incentive mechanism with bandwidth allocation for federated learning

Yingyun GUO¹^,², Bo GAO¹^,², Zhifei ZHANG¹^,², Yu ZHANG³, Ke XIONG¹^,²

¹ School of Computer and Information Technology, Beijing Jiaotong University, Beijing100044, China
² Engineering Research Center of High Speed Railway Network Management, Ministry of Education, Beijing Jiaotong University, Beijing 100044, China
³ State Grid Energy Research Institute Co., Ltd., Beijing 102209, China

Revised:2022-09-16 Online:2022-12-30 Published:2022-12-01
Supported by:
The National Natural Science Foundation of China(61872028);The Fundamental Research Funds for the Central Universities(2021JBM008);The Fundamental Research Funds for the Central Universities(2022JBXT001)

摘要/Abstract

摘要：

联邦学习（FL, federated learning）是一种新兴的机器学习范式，它可以充分利用移动众包资源进行去中心化数据训练。然而，在无线网络中部署 FL 面临网络带宽有限、移动用户自私等挑战。为了应对这些挑战，提出了一种基于带宽分配的激励机制（IMBA, incentive mechanism with bandwidth allocation）。IMBA考虑用户数据质量和计算能力的不同设计支付方案，以激励高数据质量用户贡献其计算资源，进而提升模型训练精度。通过最小化训练时间和支付成本权重和确定最佳支付与带宽分配方案，通过优化带宽分配降低训练时延。实验表明， IMBA能够有效提高训练精度，降低训练时间，并帮助服务器灵活权衡训练时间和支付报酬。

关键词: 联邦学习, 激励机制, 带宽分配, Stackelberg博弈, 训练质量

Abstract:

Federated learning (FL) is an emerging machine learning paradigm that can make full use of crowd sourced mobile resources for training on decentralized data.However, it is challenging to deploy FL over a wireless network because of the limited bandwidth and clients’ selfishness.To address these challenges, an incentive mechanism with bandwidth allocation (IMBA) was proposed.Considering the difference between clients' data quality and computing power, IMBA designs a payment scheme to incentivize high-quality clients to contribute their computing resources, thus improving the training accuracy of the model.By minimizing the weight sum of training time and payment cost, the optimal payment and bandwidth allocation scheme was determined, and the training delay was reduced by optimizing bandwidth allocation.Experiments show that IMBA effectively improves training accuracy, reduces the training delay and helps the server flexibly balance training delay and hiring payment.

Key words: federated learning, incentive mechanism, bandwidth allocation, Stackelberg game, training quality

中图分类号:

TP393

郭英芸, 高博, 张志飞, 张煜, 熊轲. 一种基于带宽分配的联邦学习激励机制[J]. 物联网学报, 2022, 6(4): 82-92.

Yingyun GUO, Bo GAO, Zhifei ZHANG, Yu ZHANG, Ke XIONG. An incentive mechanism with bandwidth allocation for federated learning[J]. Chinese Journal on Internet of Things, 2022, 6(4): 82-92.

图/表 10

表1

参数说明"

符号	说明	符号	说明
$N / N$	用户数量/用户集合	$O_{n} / O_{n}$	用户n的训练数据数量/集合
B	总上行带宽	$λ / λ_{n}$	带宽分配比例向量/用户n的带宽比例
$g_{n}$	用户n与基站间信道增益	$h_{n}$	用户n的小尺度衰落分量
$r_{n}$	用户n的上行传输速率	$ρ_{n}$	用户n的阴影衰落模型
$ζ_{n}$	用户n的路径损失模型	$Z_{0}$	噪声功率
$p_{n}$	用户n的发射功率	D	训练模型大小
$t_{n}^{M}$	用户n的通信时间，M表示通信	$E_{n}^{M}$	用户n的通信功耗
$θ_{n}^{M}$	用户n的单位通信成本	$c_{n}^{M}$	用户n的通信成本
$I_{n}$	用户n的本地epoch数	$q_{n}$	用户n的数据质量
$Δ$	单位样本的大小	$e_{n}$	用户n计算芯片组的有效电容参数
$f_{n}$	用户n用于本地训练的CPU频率	$t_{n}^{P}$	用户n本地计算时间，P表示计算
$E_{n}^{P}$	用户n的本地计算功耗	$θ_{n}^{P}$	用户n单位计算成本
$c_{n}^{P}$	用户n的本地计算成本	$t_{n}$	用户n训练时间
$c_{n}$	用户n的本地成本	$δ_{n}$	用户n的数据非独立同分布程度
$μ_{n}$	用户n处理单位比特样本数据所需的CPU周期数	$τ / τ_{n}$	服务器的单位支付向量/用户n的单位报酬
$β$	时间与支付成本的权衡系数	$K / K$	参与训练用户数量/训练用户集合

表1

图1

图2

图3

图4

图5

图6

图7

图8

表2

参考文献 31

[1]	MCMAHAN B , MOORE E , RAMAGE D ,et al. Communication-efficient learning of deep networks from decentralized data[C]// Artificial intelligence and statistics. New York:PMLR, 2017: 1273-1282.
[2]	LIM W Y B , LUONG N C , HOANG D T ,et al. Federated learning in mobile edge networks:A comprehensive survey[J]. IEEE Communications Surveys ＆ Tutorials, 2020,22(3): 2031-2063.
[3]	CHEN Y Q , QIN X , WANG J D ,et al. Fed Health:a federated transfer learning framework for wearable healthcare[J]. IEEE Intelligent Systems, 2020,35(4): 83-93.
[4]	LIU D B , DLIGACH D , MILLER T ,et al. Two-stage federated phenotyping and patient representation learning[EB]. 2019.
[5]	微众银行. 联邦学习白皮书v2.0[R]. 2020.
	WEBANK. Federated learning white paper v2.0[R]. 2020.
[6]	QIN Z , LI G Y , YE H . Federated learning and wireless communications[J]. IEEE Wireless Communications, 2021,28(5): 134-140.
[7]	SHI W Q , ZHOU S , NIU Z S . Device scheduling with fast conver gence for federated learning[C]// Proceedings of ICC 2020-2020 IEEE International Conference on Communications. Piscataway:IEEE Press, 2020: 1-6.
[8]	YANG Z H , CHEN M Z , SAAD W ,et al. Energy efficient federated learning over wireless communication networks[J]. IEEE Transactions on Wireless Communications, 2021,20(3): 1935-1949.
[9]	CHEN M Z , YANG Z H , SAAD W ,et al. A joint learning and communications framework for federated learning over wireless networks[J]. IEEE Transactions on Wireless Communications, 2021,20(1): 269-283.
[10]	REN J K , YU G D , DING G Y . Accelerating DNN training in wireless federated edge learning systems[J]. IEEE Journal on Selected Areas in Communications, 2021,39(1): 219-232.
[11]	ABDULRAHMAN S , TOUT H , MOURAD A ,et al. FedMCCS:multicriteria client selection model for optimal IoT federated learning[J]. IEEE Internet of Things Journal, 2021,8(6): 4723-4735.
[12]	ZHANG W Y , WANG X M , ZHOU P ,et al. Client selection for federated learning with non-IID data in mobile edge computing[J]. IEEE Access, 2021,9: 24462-24474.
[13]	CHEN M Z , POOR H V , SAAD W ,et al. Convergence time minimization of federated learning over wireless networks[C]// Proceedings of ICC 2020-2020 IEEE International Conference on Communications. Piscataway:IEEE Press, 2020: 1-6.
[14]	NISHIO T , YONETANI R . Client selection for federated learning with heterogeneous resources in mobile edge[C]// Proceedings of ICC 2019-2019 IEEE International Conference on Communications. Piscataway:IEEE Press, 2019: 1-7.
[15]	CHAIZ , ALI A , ZAWAD S ,et al. TiFL:a tier-based federated learning system[EB]. 2020.
[16]	ZENG Q S , DU Y Q , HUANG K B ,et al. Energy-efficient radio resource allocation for federated edge learning[C]// Proceedings of 2020 IEEE International Conference on Communications Workshops. Piscataway:IEEE Press, 2020: 1-6.
[17]	XU J , WANG H . Client selection and bandwidth allocation in wireless federated learning networks:a long-term perspective[J]. IEEE Transactions on Wireless Communications, 2020,20(2): 1188-1200.
[18]	ZHU G X , WANG Y , HUANG K B . Broadband analog aggregation for low-latency federated edge learning[J]. IEEE Transactions on Wireless Communications, 2020,19(1): 491-506.
[19]	LIM W Y B , HUANG J Q , XIONG Z H ,et al. Towards federated learning in UAV-enabled Internet of vehicles:a multi-dimensional contract-matching approach[J]. IEEE Transactions on Intelligent Transportation Systems, 2021,22(8): 5140-5154.
[20]	NISHIO T , SHINKUMA R , MANDAYAM N B . Estimation of individual device contributions for incentivizing federated learning[C]// Proceedings of 2020 IEEE Globecom Workshops (GC Wkshps. Piscataway:IEEE Press, 2020: 1-6.
[21]	WANG G , DANG C X , ZHOU Z Y . Measure contribution of participants in federated learning[C]// Proceedings of 2019 IEEE International Conference on Big Data(Big Data). Piscataway:IEEE Press, 2019: 2597-2604.
[22]	WU M Q , YE D D , DING J H ,et al. Incentivizing differentially private federated learning:a multidimensional contract approach[J]. IEEE Internet of Things Journal, 2021,8(13): 10639-10651.
[23]	PANDEY S R , TRAN N H , BENNIS M ,et al. A crowdsourcing framework for on-device federated learning[J]. IEEE Transactions on Wireless Communications, 2020,19(5): 3241-3256.
[24]	LI M , WENG J , YANG A J ,et al. CrowdBC:a blockchain-based decentralized framework for crowdsourcing[J]. IEEE Transactions on Parallel and Distributed Systems, 2019,30(6): 1251-1266.
[25]	KANG J W , XIONG Z H , NIYATO D ,et al. Incentive mechanism for reliable federated learning:a joint optimization approach to combining reputation and contract theory[J]. IEEE Internet of Things Journal, 2019,6(6): 10700-10714.
[26]	JIAO Y T , WANG P , NIYATO D ,et al. Toward an automated auction framework for wireless federated learning services market[J]. IEEETransactions on Mobile Computing, 2021,20(10): 3034-3048.
[27]	TANG M , WONG V W . An incentive mechanism for cross-silo federated learning:a public goods perspective[C]// Proceedings of the IEEE Conference on Computer Communications (INFOCOM). IEEE, 2021: 1-10.
[28]	DING N N , FANG Z X , HUANG J W . Incentive mechanism design for federated learning with multi-dimensional private information[C]// Proceedings of 2020 18th International Symposium on Modeling and Optimization in Mobile,Ad Hoc,and Wireless Networks (WiOPT). Piscataway:IEEE Press, 2020: 1-8.
[29]	SHANG L C , WANG X H , WANG P ,et al. Computation offloading management in vehicular edge network under imperfect CSI[C]// Proceedings of 2019 IEEE 2nd International Conference on Information Communication and Signal Processing. Piscataway:IEEE Press, 2019: 199-203.
[30]	ZHAO Y , LI M , LAI L Z ,et al. Federated learning withnon-IIDdata[EB]. 2018.
[31]	NARAHARI Y . 博弈论与机制设计[M]. 曹乾,译. 北京: 中国人民大学出版社, 2017.
	NARAHARI Y . Game theory and mechanism design[M]. CAO Q,Translator. Beijing: Chinese People’s Publishing House., 2017.

参数β/(×1 ¹⁰0 )	参与训练的用户序号
2，3，4	11，14，15，16，17，18，19，20
5，6	7，11，14，15，16，17，18，19，20
7，8，9，10，11，12	7，11，12，14，15，16，17，18，19，20
13	7，8，11，12，14，15，16，17，18，19，20
14	7，8，11，12，13，14，15，16，17，18，19，20
15	7，8，10，11，12，13，14，15，16，17，18，19，20

一种基于带宽分配的联邦学习激励机制

An incentive mechanism with bandwidth allocation for federated learning

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 31

相关文章 8

Metrics

推荐阅读 0

[1]	耿光磊, 高博, 熊轲, 樊平毅, 陆杨, 王煜炜. 联邦学习赋能6G网络综述[J]. 物联网学报, 2023, 7(2): 50-66.
[2]	郭佳慧, 陈卓越, 高玮, 王玺钧, 孙兴华, 高林. 基于背包模型的联邦学习客户端选择方法[J]. 物联网学报, 2022, 6(4): 158-168.
[3]	马嘉华, 孙兴华, 夏文超, 王玺钧, 谭洪舟, 朱洪波. 基于标签量信息的联邦学习节点选择算法[J]. 物联网学报, 2021, 5(4): 46-53.
[4]	石光明, 肖泳, 李莹玉, 高大化, 谢雪梅. 面向万物智联的语义通信网络[J]. 物联网学报, 2021, 5(2): 26-36.
[5]	沈晓东, 刘俊勇, 胡帅, 饶萍, 何迈. 泛在电力物联网形势下考虑多主体投资的配电网运行优化策略[J]. 物联网学报, 2021, 5(1): 117-122.
[6]	江常坤,高林. 基于双边拍卖的群智感知数据复用机制设计[J]. 物联网学报, 2019, 3(3): 26-33.
[7]	骆淑云,温雨舟,徐伟强,朱升宏. 移动智能边缘计算激励机制研究[J]. 物联网学报, 2019, 3(2): 80-88.
[8]	刘怀哲,高林. 基于移动热点共享的物联网激励机制研究[J]. 物联网学报, 2019, 3(1): 20-29.