激励跨筒仓联邦学习中的数据贡献
Incentivizing Data Contribution in Cross-Silo Federated Learning
摘要:
在跨筒仓联邦学习中,客户端(例如,组织)使用本地数据训练共享的全局模型。然而,由于隐私问题,客户端在训练期间可能不会提供足够的数据点。为了解决这个问题,我们提出了一个通用的激励框架,从全球模型中获得的利润/利益可以适当地分配给客户,以激励数据贡献。我们将客户之间的互动描述为一个数据贡献博弈,并研究其均衡。我们的特征的均衡存在的条件,并证明每个客户端的均衡数据的贡献增加其数据质量和隐私敏感性降低。我们进一步使用CIFAR-10进行实验,结果与分析一致。此外,我们发现,实际的分配机制,如线性比例,留一法,夏普利值激励更多的数据贡献,从客户端的高质量的数据,其中留一法往往达到最高的全局模型精度在平衡。
英文摘要:
In cross-silo federated learning, clients (e.g., organizations) train a shared global model using local data. However, due to privacy concerns, the clients may not contribute enough data points during training. To address this issue, we propose a general incentive framework where the profit/benefit obtained from the global model can be appropriately allocated to clients to incentivize data contribution. We formulate the clients' interactions as a data contribution game and study its equilibrium. We characterize conditions for an equilibrium to exist, and prove that each client's equilibrium data contribution increases in its data quality and decreases in the privacy sensitivity. We further conduct experiments using CIFAR-10 and show that the results are consistent with the analysis. Moreover, we show that practical allocation mechanisms such as linearly proportional, leave-one-out, and Shapley-value incentivize more data contribution from clients with higher-quality data, in which leave-one-out tends to achieve the highest global model accuracy at equilibrium.
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。