Synthetic data shuffling accelerates the convergence of federated learning under data heterogeneity
This work addresses data heterogeneity in federated learning, which is a critical problem for distributed machine learning systems, by providing a novel theoretical and practical solution that is incremental but impactful.
The paper tackles the challenge of data heterogeneity in federated learning by establishing a theoretical link between data shuffling and convergence acceleration, proving that shuffling can quadratically reduce gradient dissimilarity. It proposes a practical method using synthetic data shuffling to address privacy concerns, showing experimental improvements in performance for existing algorithms.
In federated learning, data heterogeneity is a critical challenge. A straightforward solution is to shuffle the clients' data to homogenize the distribution. However, this may violate data access rights, and how and when shuffling can accelerate the convergence of a federated optimization algorithm is not theoretically well understood. In this paper, we establish a precise and quantifiable correspondence between data heterogeneity and parameters in the convergence rate when a fraction of data is shuffled across clients. We prove that shuffling can quadratically reduce the gradient dissimilarity with respect to the shuffling percentage, accelerating convergence. Inspired by the theory, we propose a practical approach that addresses the data access rights issue by shuffling locally generated synthetic data. The experimental results show that shuffling synthetic data improves the performance of multiple existing federated learning algorithms by a large margin.