Global Convergence of Federated Learning for Mixed Regression
Global Convergence of Federated Learning for Mixed Regression
复制标题
DOI:
10.48550/arxiv.2206.07279
复制
发表时间:
2022-06
期刊:
影响因子:
--
通讯作者:
Lili Su;Jiaming Xu;Pengkun Yang
中科院分区:
文献类型:
--
作者:
Lili Su;Jiaming Xu;Pengkun Yang
This paper studies the problem of model training under Federated Learning when clients exhibit cluster structure. We contextualize this problem in mixed regression, where each client has limited local data generated from one of $k$ unknown regression models. We design an algorithm that achieves global convergence from any initialization, and works even when local data volume is highly unbalanced -- there could exist clients that contain $O(1)$ data points only. Our algorithm first runs moment descent on a few anchor clients (each with $\tilde{\Omega}(k)$ data points) to obtain coarse model estimates. Then each client alternately estimates its cluster labels and refines the model estimates based on FedAvg or FedProx. A key innovation in our analysis is a uniform estimate on the clustering errors, which we prove by bounding the VC dimension of general polynomial concept classes based on the theory of algebraic geometry.