GQ-FSL: Green Quantized Federated Split Learning
2026-07-31 • Machine Learning
Machine LearningDistributed, Parallel, and Cluster Computing
AI summaryⓘ
The authors address the problem of running big deep learning models on devices like phones that have limited battery and computing power. They propose a method called green quantized federated split learning (GQ-FSL), which reduces energy use by compressing data and computations differently on the device and the server. They also create models to understand the energy tradeoffs and ensure the learning still works well even if data varies a lot between devices. Their approach helps large neural networks run more efficiently on small devices compared to existing methods.
Deep Neural NetworksFederated LearningSplit LearningQuantizationEnergy ConsumptionEdge ComputingConvergence BoundWireless TransmissionHeterogeneous DataOptimization
Authors
Idan Roth, Lutz Lampe
Abstract
Deploying state-of-the-art deep neural networks (DNNs) at the wireless edge is severely bottlenecked by the strict energy and resource constraints of mobile devices. While federated split learning (FSL) mitigates on-device computation by offloading workloads to an edge server, this may introduce systemic overheads, while the continuous exchange of cut-layer data, and submodels still incurs significant energy consumption (EC). To address this, we propose a green quantized FSL (GQ-FSL) framework that incorporates stochastic quantization for both local collaborative training and wireless transmissions. Notably, GQ-FSL supports asymmetric precision levels for the client- and server-side submodels, effectively decoupling device energy constraints from global convergence degradation. To quantify these tradeoffs, we develop parameterized energy models for the split architecture and derive a theoretical convergence bound under statistically heterogeneous data. Building on that, we formulate a joint optimization problem to configure the DNN split point and precision levels, minimizing the total system EC while satisfying a strict target accuracy constraint. Ultimately, we demonstrate that GQ-FSL enables large-scale DNN deployment on resource-constrained devices, achieving superior energy efficiency compared to quantized federated learning and full-precision FSL.