Incentive Mechanism Design for Unbiased Federated Learning with Randomized Client Participation
Incentive Mechanism Design for Unbiased Federated Learning with Randomized Client Participation
复制标题
DOI：
10.1109/icdcs57875.2023.00027
复制
发表时间：
2023-04
期刊：
影响因子：
--
通讯作者：
Bing Luo;Yutong Feng;Shiqiang Wang;Jianwei Huang;L. Tassiulas
中科院分区：
文献类型：
--
作者：
Bing Luo;Yutong Feng;Shiqiang Wang;Jianwei Huang;L. Tassiulas
当理性客户端与服务器端在全局模型中没有相同的兴趣时，激励机制对联邦学习（FL）至关重要。然而，由于系统异构性和有限的预算，服务器激励所有客户端参与所有培训轮（称为完全参与）通常是不切实际的。现有的FL激励机制通常是通过基于客户端的数据量或系统资源来激励固定子集的客户端来设计的。因此，FL在整个训练过程中只使用这部分客户端进行，由于数据的异质性，导致模型有偏差。本文提出了一种基于博弈论的客户随机参与FL激励机制，服务器采用定制的定价策略，激励不同的客户以不同的参与水平（概率）加入FL，以获得一个无偏的高性能模型。每个客户端通过选择其最佳参与水平来响应服务器的货币激励，以最大化其利润，不仅基于所产生的本地成本，而且基于其对于全局模型的内在价值。为了有效地评估客户端对模型性能的贡献，我们推导出一个新的收敛界，分析预测客户端的任意参与水平和他们的异构数据如何影响模型的性能。通过求解一个非凸优化问题，我们的分析表明，内在价值导致服务器和客户端之间的双向支付的有趣的可能性。在硬件原型上使用真实的数据集的实验结果表明，我们的机制在实现更高的模型性能的服务器，以及更高的利润为客户端的优越性。
Incentive mechanism is crucial for federated learning (FL) when rational clients do not have the same interests in the global model as the server. However, due to system heterogeneity and limited budget, it is generally impractical for the server to incentivize all clients to participate in all training rounds (known as full participation). The existing FL incentive mechanisms are typically designed by stimulating a fixed subset of clients based on their data quantity or system resources. Hence, FL is performed only using this subset of clients throughout the entire training process, leading to a biased model because of data heterogeneity. This paper proposes a game-theoretic incentive mechanism for FL with randomized client participation, where the server adopts a customized pricing strategy that motivates different clients to join with different participation levels (probabilities) for obtaining an unbiased and high-performance model. Each client responds to the server's monetary incentive by choosing its best participation level, to maximize its profit based on not only the incurred local cost but also its intrinsic value for the global model. To effectively evaluate clients' contribution to the model performance, we derive a new convergence bound which analytically predicts how clients' arbitrary participation levels and their heterogeneous data affect the model performance. By solving a non-convex optimization problem, our analysis reveals that the intrinsic value leads to the interesting possibility of bi-directional payment between the server and clients. Experimental results using real datasets on a hardware prototype demonstrate the superiority of our mechanism in achieving higher model performance for the server as well as higher profits for the clients.