Supplementary File: Coded Cooperative Networks for Semi-Decentralized Federated Learning

TL;DR

This paper introduces a deterministic coded network approach for semi-decentralized federated learning that leverages wireless diversity, improving resilience to stragglers without needing prior network information.

Contribution

It proposes a novel deterministic coded network scheme for semi-decentralized FL that does not require prior network topology knowledge and analyzes its outage and convergence properties.

Findings

The proposed method outperforms benchmark schemes in simulations.

The scheme effectively leverages wireless diversity for improved resilience.

Theoretical analysis confirms convergence and outage performance.

Abstract

To enhance straggler resilience in federated learning (FL) systems, a semi-decentralized approach has been recently proposed, enabling collaboration between clients. Unlike the existing semi-decentralized schemes, which adaptively adjust the collaboration weight according to the network topology, this letter proposes a deterministic coded network that leverages wireless diversity for semi-decentralized FL without requiring prior information about the entire network. Furthermore, the theoretical analyses of the outage and the convergence rate of the proposed scheme are provided. Finally, the superiority of our proposed method over benchmark methods is demonstrated through comprehensive simulations.