N Zhang, M Tao - 2020 IEEE International Conference on …, 2020 - ieeexplore.ieee.org
… over-the-air federatedlearning over fading channels by taking the gradient distribution into account. The structure of the optimal power control for SGD learning mainly depends on the …
Y Li, W Zhou, H Wang, H Mi, TM Hospedales - arXiv preprint arXiv …, 2021 - arxiv.org
… being desired to maximise performance under heterogeneous data statistics across peers. … data statistics across peers we introduce a new optimization strategy to find the gradient …
W Liu, L Chen, Y Chen, W Zhang - IEEE Transactions on …, 2020 - ieeexplore.ieee.org
… Motivated by the above observations, we propose a new federatedlearning design of Momentum FederatedLearning (MFL) in this paper. In the proposed MFL design, we introduce …
S Wu, D Huang, H Wang - … of Business & Economic Statistics, 2023 - Taylor & Francis
… We study a fully decentralized federatedlearning algorithm, which is a novel gradient … it as a network gradient descent (NGD) method. In the NGD method, only statistics (eg, parameter …
… We propose a new federatedlearning algorithm, FedPAGE, able to further reduce the … than previous local methods for both federated convex and nonconvex optimization. Concretely, 1) …
… In this section, we propose a generalized version of the local stochastic gradient descent (SGD) method for federatedlearning which uses compressed signals to reduce the overall …
A Mitra, R Jaafar, GJ Pappas… - Advances in Neural …, 2021 - proceedings.neurips.cc
… federatedlearning (FL) setup where a group of clients periodically coordinate with a central server to train a statistical … convergence rates under aggressive gradient sparsification, and …
Y Bao, M Crawshaw, S Luo… - … on Machine Learning, 2022 - proceedings.mlr.press
As a prevalent distributed learning paradigm, FederatedLearning (FL) trains a global model on a massive amount of devices with infrequent communication. This paper investigates a …
… to learn from extremely large number of devices in federated setting… efficient algorithms for federatedlearning with provable … of local Gradient Descent (GD) and local Stochastic Gradient …