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

TL;DR

MAR-FL tackles scalable, communication-efficient P2P Federated Learning for wireless networks by introducing iterative group-based aggregation (MAR) with DHT coordination. It achieves per-iteration communication complexity of $O(N log N)$ and supports optional KD and DP. Empirical results on MNIST and 20NG show MAR-FL matches centralized FedAvg and other P2P baselines in accuracy while significantly reducing communication; MKD further reduces total communication, with DP enabling privacy-utility trade-offs. The work underscores MAR-FL as a practical foundation for decentralized FL in next-gen wireless settings, with future work on partial participation, approximate aggregation, and DP integration.

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.

Figures (25)

• Figure 1: MAR-FL (for $i$-th peer)
• Figure 2: Moshpit-KD (for $i$-th peer in MKD round $g$ of FL iteration $t$)
• Figure 3: MNIST
• Figure 4: 20NG
• Figure 7: Model Performance -- MAR-FL -- 20NG