Decentralized Aggregation for Energy-Efficient Federated Learning via Overlapped Clustering and D2D Communications

TL;DR

This paper introduces a decentralized federated learning scheme called FL-EOCD that uses overlapped clustering and D2D communications to reduce energy consumption and latency while maintaining convergence, eliminating the need for a central server.

Contribution

The paper proposes a novel decentralized FL scheme leveraging overlapped clustering and D2D communications, with a joint optimization for energy efficiency and convergence.

Findings

FL-EOCD reduces energy consumption compared to baseline schemes.

The scheme maintains comparable convergence rates.

Simulations show improved latency and energy efficiency.

Abstract

Federated learning (FL) has emerged as a distributed machine learning (ML) technique to train models without sharing users' private data. In this paper, we propose a decentralized FL scheme that is called \underline{f}ederated \underline{l}earning \underline{e}mpowered \underline{o}verlapped \underline{c}lustering for \underline{d}ecentralized aggregation (FL-EOCD). The introduced FL-EOCD leverages device-to-device (D2D) communications and overlapped clustering to enable decentralized aggregation, where a cluster is defined as a coverage zone of a typical device. The devices located on the overlapped clusters are called bridge devices (BDs). In the proposed FL-EOCD scheme, a clustering topology is envisioned where clusters are connected through BDs, so as the aggregated models of each cluster is disseminated to the other clusters in a decentralized manner without the need for a global aggregator or an additional hop of transmission. Unlike the star-based FL, the proposed FL-EOCD scheme involves a large number of local devices by reusing the RRBs in different non-adjacent clusters. To evaluate our proposed FL-EOCD scheme as opposed to baseline FL schemes, we consider minimizing the overall energy-consumption of devices while maintaining the convergence rate of FL subject to its time constraint. To this end, a joint optimization problem, considering scheduling the local devices/BDs to the CHs and computation frequency allocation, is formulated, where an iterative solution to this joint problem is devised. Extensive simulations are conducted to verify the effectiveness of the proposed FL-EOCD algorithm over FL conventional schemes in terms of energy consumption, latency, and convergence rate.