FedAvgMServerOptimizer<T>

Models & Types · AiDotNet.FederatedLearning.ServerOptimizers

FedAvgM server optimizer — server-side momentum for stabilized federated averaging.

Inheritance: FederatedLearningComponentBase<T> → FederatedServerOptimizerBase<T> → FedAvgMServerOptimizer<T>

For Beginners

Momentum helps smooth updates across rounds. Instead of applying only the current round’s update, the server maintains a running “velocity” that accumulates updates. This reduces oscillations caused by heterogeneous client data.

How It Works

Reference: Hsu, T.-M. H., et al. (2019). “Measuring the Effects of Non-Identical Data Distribution for Federated Visual Classification.” NeurIPS Workshop 2019.

Constructors

Constructor Summary
FedAvgMServerOptimizer(double learningRate, double momentum)

Methods

Method Summary
Vector<T> Step(Vector<T> currentGlobalParameters, Vector<T> aggregatedTargetParameters)
string GetOptimizerName()