QuantFL: Sustainable Federated Learning for Edge IoT via Pre-Trained Model Quantisation

TL;DR

QuantFL introduces a sustainable federated learning framework that leverages pre-trained models and aggressive quantisation to significantly reduce communication energy costs on IoT devices while maintaining high accuracy.

Contribution

The paper proposes QuantFL, a novel federated learning approach using pre-trained models and memory-efficient quantisation to lower energy consumption and communication costs on edge IoT devices.

Findings

Reduces total communication by 40% on MNIST and CIFAR-100.

Achieves over 80% uplink quantisation with maintained accuracy.

Delivers high accuracy with significantly fewer bits, enabling scalable IoT training.

Abstract

Federated Learning (FL) enables privacy-preserving intelligence on Internet of Things (IoT) devices but incurs a significant carbon footprint due to the high energy cost of frequent uplink transmission. While pre-trained models are increasingly available on edge devices, their potential to reduce the energy overhead of fine-tuning remains underexplored. In this work, we propose QuantFL, a sustainable FL framework that leverages pre-trained initialisation to enable aggressive, computationally lightweight quantisation. We demonstrate that pre-training naturally concentrates update statistics, allowing us to use memory-efficient bucket quantisation without the energy-intensive overhead of complex error-feedback mechanisms. On MNIST and CIFAR-100, QuantFL reduces total communication by 40\% ($\simeq40\%$ total-bit reduction with full-precision downlink; $\geq80\%$ on uplink or when downlink is quantised) while matching or exceeding uncompressed baselines under strict bandwidth budgets; BU attains 89.00\% (MNIST) and 66.89\% (CIFAR-100) test accuracy with orders of magnitude fewer bits. We also account for uplink and downlink costs and provide ablations on quantisation levels and initialisation. QuantFL delivers a practical, "green" recipe for scalable training on battery-constrained IoT networks.