… control with unknown gradient statistics: We propose an adaptive power control algorithm
that estimates the gradient statistics based on the historical aggregated gradients and then …

… over-the-air federated learning over fading channels by taking the gradient distribution into
account. The structure of the optimal power control for SGD learning mainly depends on the …

… 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 …

… Motivated by the above observations, we propose a new federated learning design of
Momentum Federated Learning (MFL) in this paper. In the proposed MFL design, we introduce …

… We study a fully decentralized federated learning 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 federated learning 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 federated learning which uses compressed signals to reduce the overall …

… federated learning (FL) setup where a group of clients periodically coordinate with a central
server to train a statistical … convergence rates under aggressive gradient sparsification, and …

As a prevalent distributed learning paradigm, Federated Learning (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
federated learning with provable … of local Gradient Descent (GD) and local Stochastic Gradient …