发表机构
Amar Nath and Shashi Khosla School of Information Technology, Indian Institute of Technology Delhi; Department of Electrical Engineering, Indian Institute of Technology Delhi; College of Computing and Data Science, Nanyang Technological University（印度理工学院德里分校 Amar Nath and Shashi Khosla 信息技术学院; 印度理工学院德里分校电气工程系; 南洋理工大学计算与数据科学学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对联邦学习类不平衡问题，提出DAFL分布式增强框架，通过理论建立增强与收敛关系并联合优化增强量和训练时间，提升少数类F1分数并显著减少训练时间。
AI 中文摘要
在联邦学习中，缓解类不平衡对于提升少数类性能至关重要。解决该问题的一种常见方法是对少数类样本进行增强，以实现局部类平衡。现有方法将增强视为启发式手段，并未建立增强量如何影响联邦学习收敛性的理论，导致过度增强和训练时间增加。为解决此局限，我们首先建立了增强与联邦学习收敛行为之间的关系。利用这一洞察，我们提出了DAFL，一种分布式增强框架，通过联合最小化增强量和训练时间，同时约束全局类不平衡，来确定每个客户端-类对所需的最小增强量，从而提升少数类F1分数。实验结果表明，DAFL在显著减少训练时间的同时持续提升少数类F1分数，尤其是在严重的全局类不平衡和高标签比例不平衡情况下。
英文摘要
In federated learning, mitigating class imbalance is essential to improve minority-class performance. A common approach to address this problem is to augment minority-class samples to achieve local class balance. Existing approaches treat augmentation as a heuristic and do not establish how the amount of augmentation influences the convergence of federated learning, leading to excessive augmentation and increased training time. To address this limitation, we first establish the relationship between augmentation and the convergence behavior of federated learning. Leveraging this insight, we propose DAFL, a distributed augmentation framework that determines the minimum augmentation required for each client-class pair by jointly minimizing augmentation and training time while constraining global class imbalance, thereby improving minority-class F1-score. Experimental results demonstrate that DAFL consistently improves minority-class F1-score while substantially reducing training time, particularly under severe global class imbalance and high label proportion imbalance.
Comments8 pages
