---
type: "Research"
title: "FedAvg — McMahan et al. (2017)"
description: "Communication-efficient decentralised training by model averaging."
resource: "https://arxiv.org/abs/1602.05629"
tags: [fl, foundational, research]
generated: { by: human:crpage, at: 2026-07-09T09:44:00Z }
status: stable
sources: [{ id: primary, resource: "https://arxiv.org/abs/1602.05629" }]
---

Established the core FL training loop: iterative client-side training with **model averaging** dramatically reduces communication relative to synchronised SGD, while handling non-IID and unbalanced data.

Relates to: [Federated learning](../federated/federated-learning.md).