Answers are grounded in this paper and cite their sources.
Running Federated Learning (FL) on resource-constrained devices is challenging due to the resources required for training. Layer freezing has been proposed to reduce the computational costs and thus accelerate training. However, we identify that existing layer freezing approaches either learn quickly or learn effectively, but do not balance them. Specifically, aggressive early-stage layer freezing accelerates training but achieves a lower final accuracy. On the other hand, accuracy-guaranteed layer freezing obtains higher final accuracy but with marginal training time improvement. This article proposes FedFreeze – a dual-phase layer freezing federated learning framework that, for the first time, combines early-stage and accuracy-guaranteed layer freezing into a unified mechanism. FedFreeze designs a novel regularization-based layer freezing strategy on the device to apply early-stage layer freezing even during the initial stages for improving training speedup. In addition, FedFreeze develops a convergence-based layer freezing strategy to achieve a high final accuracy. Experimental results demonstrate that the FedFreeze framework achieves training speedup of up to 1.3× while maintaining a high final accuracy with an accuracy loss of no more than 1.67% compared to vanilla FL. Compared to state-of-the-art early-stage and accuracy-guaranteed layer freezing approaches, FedFreeze balances accuracy and speedup performance across different experimental testbeds, datasets, models, and initialization methods.
No takes yet. Share an insight, caveat, or question.
Wu et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: