Federated User Clustering for non-IID Federated Learning

Abstract

Federated Learning (FL) is one of the leading learning paradigms for enabling a more significant presence of intelligent applications in networking considering highly distributed environments while preserving user privacy.
However, FL has the significant shortcoming of requiring user data to be Independent Identically Distributed (IID) to make reliable predictions for a given group of users.
We present a Neural Network-based Federated Clustering mechanism capable of clustering the local models trained by users of the network with no access to their raw data.
We also present an alternative to the FedAvg aggregation algorithm used in traditional FL, which significantly increases the aggregated models' reliability in Mean Square Error by creating several training models over IID users.