Federated Learning via Plurality Vote
This addresses efficiency and security challenges in federated learning for edge devices, but it is incremental as it builds on existing quantization and voting techniques.
The paper tackles the joint optimization of communication overhead, learning reliability, and deployment efficiency in federated learning by proposing FedVote, which uses binary or ternary weights to reduce quantization error and achieve faster convergence compared to direct quantization methods.
Federated learning allows collaborative workers to solve a machine learning problem while preserving data privacy. Recent studies have tackled various challenges in federated learning, but the joint optimization of communication overhead, learning reliability, and deployment efficiency is still an open problem. To this end, we propose a new scheme named federated learning via plurality vote (FedVote). In each communication round of FedVote, workers transmit binary or ternary weights to the server with low communication overhead. The model parameters are aggregated via weighted voting to enhance the resilience against Byzantine attacks. When deployed for inference, the model with binary or ternary weights is resource-friendly to edge devices. We show that our proposed method can reduce quantization error and converges faster compared with the methods directly quantizing the model updates.