预算约束下多任务联邦学习激励机制

doi:10.19678/j.issn.1000-3428.0067742

摘要/Abstract

摘要： 联邦学习是一种实现数据隐私保护的分布式机器学习范式,性能取决于数据源的质量和数据规模。客户端是理性个体,参与联邦学习将耗费计算、通信和隐私等成本,需要通过激励提高客户端的参与意愿。因此联邦学习能成功应用的关键之一是尽可能多地激励高质量数据客户端参与训练。多任务联邦学习环境下客户端拥有面向不同任务且质量不同的数据,并具有执行能力的约束。为提高多个学习任务的整体性能,在预算受限的条件下设计一种面向任务的客户选择和报酬机制。通过分析影响模型精度的重要因素,提出一种基于客户端数据样本分布特征的质量评估标准,并结合客户端成本信息,设计一种逆向拍卖的激励机制(EMD-MQMFL),实现客户端的任务指派和支付策略。从理论上分析和证明了该机制具有诚实性、个人理性以及预算可行性,并通过大量实验验证了该方法在联邦学习性能上的有效性。在MNIST、Fashion-MNIST、Cifar-10数据集上的实验结果表明, EMD-MQMFL在数据不平衡的情况下,平均模型精度比已有的机制至少提高5.6个百分点。

关键词: 联邦学习, 多任务, 逆向拍卖, 激励机制, 数据质量

Abstract: Federated Learning (FL) is a distributed machine-learning paradigm that achieves data-privacy protection, and its performance depends on the quality and scale of the data source. The client is a rational individual, and the client's participation in FL incurs costs related to computation, communication, and privacy. Thus, the client must be encouraged to participate through incentives. One of the key factors affecting the successful application of FL is the participation of clients with high-quality data in training. In a multi-task FL environment, clients possess data that are specific to different tasks of varying quality, and their execution capabilities are limited. To improve the overall performance of multiple learning tasks, a task-oriented customer selection and a reward mechanism are designed under budget constraints in this study. By analyzing the important factors that affect the accuracy of the proposed model, a quality-evaluation standard based on the distribution characteristics of client data samples is proposed. Combining this with the client's cost information, an incentive mechanism for reverse auction (EMD-MQMFL) is designed to achieve task assignment and payment strategies for the client. This mechanism has been theoretically analyzed and proven to exhibit honesty, personal rationality, and budget feasibility. Furthermore, its effectiveness in FL performance has been verified via numerous experiments. Experimental results on the MNIST, Fashion-MNIST, and Cifar-10 datasets show that EMD-MQMFL improves the average model accuracy by at least 5.6 percentage points compared with existing mechanisms for cases involving imbalanced data.

Key words: Federated Learning(FL), multi-task, reverse auction, incentive mechanism, data quality

中图分类号:

TP391

顾永跟, 李国笑, 吴小红, 陶杰, 张艳琼. 预算约束下多任务联邦学习激励机制[J]. 计算机工程, 2024, 50(5): 149-157.

GU Yonggen, LI Guoxiao, WU Xiaohong, TAO Jie, ZHANG Yanqiong. Incentive Mechanism for Multi-Task Federated Learning Under Budget Constraints[J]. Computer Engineering, 2024, 50(5): 149-157.

https://www.ecice06.com/CN/Y2024/V50/I5/149

参考文献

[1] NGUYEN G, DLUGOLINSKY S, BOBÁK M, et al. Machine learning and deep learning frameworks and libraries for large-scale data mining:a survey[J]. Artificial Intelligence Review, 2019, 52:77-124.
[2] CHEN X J, BIAN J. Analysis of the application of machine learning algorithms in data mining[J]. Journal of Physics:Conference Series, 2020, 1550(3):e032039.
[3] WARING J, LINDVALL C, UMETON R. Automated machine learning:review of the state-of-the-art and opportunities for healthcare[J]. Artificial Intelligence in Medicine, 2020, 104:101822.
[4] XU X Y. Machine learning-based prediction of urban soil environment and corpus translation teaching[J]. Arabian Journal of Geosciences, 2021, 14:1-15.
[5] ALMASKATI N. Machine learning in finance:major applications, issues, metrics, and future trends[J]. International Journal of Financial Engineering, 2022,9(3):2250010.
[6] KIM J Y, HA H, CHUN B G, et al. Collaborative analytics for data silos[C]//Proceedings of the 32nd International Conference on Data Engineering. Washington D. C., USA:IEEE Press, 2016:743-754.
[7] 刘艺璇, 陈红, 刘宇涵, 等. 联邦学习中的隐私保护技术[J]. 软件学报, 2021, 33(3):1057-1092. LIU Y X, CHEN H, LIU Y H, et al. Privacy-preserving techniques in federated learning[J]. Journal of Software, 2021, 33(3):1057-1092.(in Chinese)
[8] REGULATION G D P. General data protection regulation (GDPR)[J]. Intersoft Consulting, Accessed in October, 2018, 24(1):1-10.
[9] MCMAHAN H B, MOORE E, RAMAGE D, et al. Communication-efficient learning of deep networks from decentralized data[EB/OL].[2023-04-25]. https://arxiv.org/pdf/1602.05629.pdf.
[10] SONG T S, TONG Y X, WEI S Y. Profit allocation for federated learning[C]//Proceedings of International Conference on Big Data. Washington D. C., USA:IEEE Press, 2019:2577-2586.
[11] ZHAN Y F, LI P, WANG K, et al. Big data analytics by crowdlearning:architecture and mechanism design[J]. IEEE Network, 2020, 34 (3):143-147.
[12] 王鑫, 周泽宝, 余芸, 等.一种面向电能量数据的联邦学习可靠性激励机制[J].计算机科学,2022,49(3):31-38. WANG X, ZHOU Z B, YU Y, et al. Reliable incentive mechanism for federated learning of electric metering data[J]. Computer Science, 2022,49(3):31-38.(in Chinese)
[13] 温依霖, 赵乃良, 曾艳, 等. 基于本地模型质量的客户端选择方法[J].计算机工程,2023,49(6):131-143. WEN Y L, ZHAO N L, ZENG Y, et al. Client selection method based on local model quality[J]. Computer Engineering, 2023,49(6):131-143. (in Chinese)
[14] 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.
[15] SARIKAYA Y, ERCETIN O. Motivating workers in federated learning:a Stackelberg game perspective[J]. IEEE Networking Letters, 2019, 2(1):23-27.
[16] 周全兴, 李秋贤, 丁红发, 等.基于博弈论优化的高效联邦学习方案[J].计算机工程,2022,48(8):144-151,159. ZHOU Q X, LI Q X, DING H F, et al. Efficient federated learning scheme based on game theory optimization[J]. Computer Engineering, 2022,48(8):144-151,159. (in Chinese)
[17] 梁文雅, 刘波, 林伟伟, 等.联邦学习激励机制研究综述[J].计算机科学,2022,49(12):46-52. LIANG W Y, LIN B, LIN W W, et al. Survey of incentive mechanism for federated learning[J]. Computer Science, 2022,49(12):46-52. (in Chinese)
[18] SMITH V, CHIANG C K, SANJABI M, et al. Federated multi-task learning[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York, USA:ACM Press, 2017:4427-4437.
[19] MARFOQ O, NEGLIA G, BELLET A, et al. Federated multi-task learning under a mixture of distributions[EB/OL].[2023-04-25]. https://arxiv.org/abs/2108.10252v4.
[20] DENG Y H, LYU F, REN J, et al. FAIR:quality-aware federated learning with precise user incentive and model aggregation[C]//Proceedings of Conference on Computer Communications. Washington D. C., USA:IEEE Press, 2021:1-10.
[21] HUI D, ZHUO L, XIN C. Quality-aware incentive mechanism design based on matching game for hierarchical federated learning[C]//Proceedings of Conference on Computer Communications Workshops. Washington D. C., USA:IEEE Press, 2022:1-6.
[22] ZHAO Y, LI M, LAI L, et al. Federated learning with non-iid data[EB/OL].[2023-04-25].http://arXiv preprint arXiv:1806.00582, 2018.
[23] 顾永跟,钟浩天,吴小红,等. 不平衡数据下预算限制的联邦学习激励机制[J].计算机应用研究,2022,39(11):3385-3389. GU Y G, ZHONG H T, WU X H, et al. Incentive mechanism for federated learning with budget constraints under unbalanced data[J]. Application Research of Computers, 2022,39(11):3385-3389.(in Chinese)
[24] MYERSON R B. Optimal auction design[J]. Mathematics of Operations Research, 1981, 6(1):58-73.
[25] LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11):2278-2324.
[26] KRIZHEVSKY A, HINTON G. Learning multiple layers of features from tiny images[J]. Handbook of Systemic Autoimmune Diseases, 2009, 1 (4):1-58.
[27] DESAI H B, OZDAYI M S, KANTARCIOGLU M. BlockFLA:accountable federated learning via hybrid blockchain architecture[C]//Proceedings of the 11th Conference on Data and Application Security and Privacy. New York, USA:ACM Press, 2021:101-112.

选择文件类型/文献管理软件名称

选择包含的内容