arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2408.04963cs.LG
收藏

LiD-FL:迈向列表可解码的联邦学习

LiD-FL: Towards List-Decodable Federated Learning

College of Computer Science, Sichuan University(四川大学计算机学院);Toyota Technological Institute at Chicago(芝加哥丰田理工学院);School of Statistics and Data Science, Nankai University(南开大学统计与数据科学学院)

Hong Liu, Liren Shan, Han Bao, Ronghui You, Yuhao Yi, Jiancheng Lv

AI总结:

本文提出列表可解码联邦学习框架,允许服务器维护模型列表,在恶意参与者过半时仍保证至少一个模型性能良好,并证明收敛性及实验验证其抗攻击能力。

摘要:

联邦学习常用于包含许多未经核实参与者的环境中。因此,在对抗性攻击下的联邦学习受到了广泛关注。本文提出了一种用于列表可解码联邦学习的算法框架,其中中央服务器维护一个模型列表,至少有一个模型保证表现良好。该框架对诚实工作者的比例没有严格限制,将拜占庭联邦学习的适用性扩展到恶意参与者超过一半的场景。在对损失函数进行适当假设的条件下,我们证明了该方法的收敛定理。实验结果(包括使用凸损失和非凸损失的图像分类任务)表明,所提出的算法能够在各种攻击下抵御恶意多数。

英文摘要:

Federated learning is often used in environments with many unverified participants. Therefore, federated learning under adversarial attacks receives significant attention. This paper proposes an algorithmic framework for list-decodable federated learning, where a central server maintains a list of models, with at least one guaranteed to perform well. The framework has no strict restriction on the fraction of honest workers, extending the applicability of Byzantine federated learning to the scenario with more than half adversaries. Under proper assumptions on the loss function, we prove a convergence theorem for our method. Experimental results, including image classification tasks with both convex and non-convex losses, demonstrate that the proposed algorithm can withstand the malicious majority under various attacks.

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。

↑