arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2108.02741cs.LGcs.DC
收藏

GIFAIR-FL:联邦学习中实现群体与个体公平性的框架

GIFAIR-FL: A Framework for Group and Individual Fairness in Federated Learning

University of Michigan(密歇根大学);American University of Beirut(贝鲁特美国大学)

Xubo Yue, Maher Nouiehed, Raed Al Kontar

AI总结:

本文提出GIFAIR-FL框架,通过正则化项惩罚客户端群体损失分散程度,在联邦学习中同时实现群体与个体公平性,并在保证收敛性和预测精度的同时提升了公平性。

摘要:

本文提出GIFAIR-FL:一个在联邦学习(Federated Learning)设置中施加群体和个体公平性的框架。通过添加正则化项,我们的算法惩罚客户端群体损失的分散程度,以驱动优化器找到公平的解决方案。我们的框架GIFAIR-FL能适应全局和个性化设置。在理论上,我们证明了在非凸和强凸设置下的收敛性。收敛保证对i.i.d.和非i.i.d.数据均成立。为验证算法的实证表现,我们将该方法应用于图像分类和文本预测任务。与现有算法相比,在保持相当或更优预测精度的同时,本方法展现出更好的公平性结果。

英文摘要:

In this paper we propose \texttt{GIFAIR-FL}: a framework that imposes \textbf{G}roup and \textbf{I}ndividual \textbf{FAIR}ness to \textbf{F}ederated \textbf{L}earning settings. By adding a regularization term, our algorithm penalizes the spread in the loss of client groups to drive the optimizer to fair solutions. Our framework \texttt{GIFAIR-FL} can accommodate both global and personalized settings. Theoretically, we show convergence in non-convex and strongly convex settings. Our convergence guarantees hold for both $i.i.d.$ and non-$i.i.d.$ data. To demonstrate the empirical performance of our algorithm, we apply our method to image classification and text prediction tasks. Compared to existing algorithms, our method shows improved fairness results while retaining superior or similar prediction accuracy.

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。

↑