Preprint
Aug 2026
Resilient Decentralized Wireless Federated Learning via Gradient Tracking with AdamW
Experimental results on heterogeneous MNIST and CIFAR-10 settings show that QEF-GT-AdamW consistently improves robustness and convergence performance over representative DecL baselines while achieving favorable accuracy-communication trade-offs under limited wireless resources.
Thieu Van Nguyen, T. Nguyen, Ons Aouedi et al.
· 0 citations