MAR-FL: A Communication Efficient Peer-to-Peer Federated Learning System

Felix Mulitze, Herbert Woisetschläger, Hans Arno Jacobsen
12/4/2025
cs.LGcs.AI

No official code found. Best implementation: felix-fjm/mar-fl (2 stars, last commit 2026-04-20).

Abstract

The convergence of next-generation wireless systems and distributed Machine Learning (ML) demands Federated Learning (FL) methods that remain efficient and robust with wireless connected peers and under network churn. Peer-to-peer (P2P) FL removes the bottleneck of a central coordinator, but existing approaches suffer from excessive communication complexity, limiting their scalability in practice. We introduce MAR-FL, a novel P2P FL system that leverages iterative group-based aggregation to substantially reduce communication overhead while retaining resilience to churn. MAR-FL achieves communication costs that scale as O(N log N), contrasting with the O(N^2) complexity of previously existing baselines, and thereby maintains effectiveness especially as the number of peers in an aggregation round grows. The system is robust towards unreliable FL clients and can integrate private computing.

View on arXivView PDF

Code Implementations(1)

felix-fjm/mar-flCommunity69%

MAR-FL: A Communication Efficient Peer-to-Peer Federated Learning System

20PythonDec 18, 20256 months ago

Similar Papers

OCD-FL: A Novel Communication-Efficient Peer Selection-based Decentralized Federated Learning

87% match2024

Exploring the Practicality of Federated Learning: A Survey Towards the Communication Perspective

86% match2024

Towards Scalable Wireless Federated Learning: Challenges and Solutions

86% match2023

FedBWO: Enhancing Communication Efficiency in Federated Learning

85% match2025

FedDec: Peer-to-peer Aided Federated Learning

85% match2023

Distributed Learning in Heterogeneous Environment: federated learning with adaptive aggregation and computation reduction

85% match2023

Cite this paper

@article{mulitze2025marfl,
  title  = {MAR-FL: A Communication Efficient Peer-to-Peer Federated Learning System},
  author = {Felix Mulitze and Herbert Woisetschläger and Hans Arno Jacobsen},
  year   = {2025},
  eprint = {2512.05234},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2512.05234v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2512.05234.svg)](https://codeofpaper.com/paper/2512.05234)

Discussion