Add to Chrome
Add to Firefox
Add to EdgeCommunication-Efficient Heterogeneous Federated Learning with Generalized Heavy-Ball Momentum
Nov 30, 2023
Federated Learning (FL) is the state-of-the-art approach for learning from decentralized data in privacy-constrained scenarios. As the current literature reports, the main problems associated with FL refer to system and statistical challenges: the former ones demand for efficient learning from edge devices, including lowering communication bandwidth and frequency, while the latter require algorithms robust to non-iidness. State-of-art approaches either guarantee convergence at increased communication cost or are not sufficiently robust to handle extreme heterogeneous local distributions. In this work we propose a novel generalization of the heavy-ball momentum, and present FedHBM to effectively address statistical heterogeneity in FL without introducing any communication overhead. We conduct extensive experimentation on common FL vision and NLP datasets, showing that our FedHBM algorithm empirically yields better model quality and higher convergence speed w.r.t. the state-of-art, especially in pathological non-iid scenarios. While being designed for cross-silo settings, we show how FedHBM is applicable in moderate-to-high cross-device scenarios, and how good model initializations (e.g. pre-training) can be exploited for prompt acceleration. Extended experimentation on large-scale real-world federated datasets further corroborates the effectiveness of our approach for real-world FL applications.
