GPFL: A Gradient Projection-Based Client Selection Framework for Efficient Federated Learning
Shijie Na, Yuzhi Liang, Siu-Ming Yiu
TL;DR
GPFL tackles the challenge of client selection in federated learning under non-IID data by introducing Gradient Projection (GP) to quantify how well a client's local gradient aligns with the global descent direction, augmented by Gradient Projection Confidence Bound (GPCB) to balance exploration and exploitation. A pre-selection strategy reduces communication and computation by filtering candidates before training, while a momentum-based gradient approach stabilizes updates. Theoretical regret bounds and convergence analysis accompany extensive experiments on FEMNIST and CIFAR-10, showing GPFL achieves faster convergence and higher test accuracy than baselines, particularly in non-IID settings, with notable improvements on FEMNIST. The framework demonstrates practical impact by lowering training time and improving robustness to data heterogeneity, paving the way for more efficient, privacy-preserving federated systems.
Abstract
Federated learning client selection is crucial for determining participant clients while balancing model accuracy and communication efficiency. Existing methods have limitations in handling data heterogeneity, computational burdens, and independent client treatment. To address these challenges, we propose GPFL, which measures client value by comparing local and global descent directions. We also employ an Exploit-Explore mechanism to enhance performance. Experimental results on FEMINST and CIFAR-10 datasets demonstrate that GPFL outperforms baselines in Non-IID scenarios, achieving over 9\% improvement in FEMINST test accuracy. Moreover, GPFL exhibits shorter computation times through pre-selection and parameter reuse in federated learning.