Enhancing Communication Efficiency in FL with Adaptive Gradient Quantization and Communication Frequency Optimization

2025-09-27

AI Summary

To address the high communication overhead caused by frequent model updates in resource-constrained wireless federated learning, this paper proposes a triple-cooperative optimization framework: (1) adaptive feature pruning to reduce local computation and upload dimensionality; (2) a dynamic quantization mechanism guided by gradient novelty and error sensitivity, enabling high-fidelity gradient compression; and (3) joint optimization of communication frequency to balance convergence speed and transmission cost. Unlike conventional static quantization and fixed communication intervals, our approach adaptively coordinates computation, compression, and scheduling. Experimental results demonstrate that, while preserving model accuracy, the proposed method reduces communication overhead by up to 42.6%, decreases required convergence rounds by 31.8%, and significantly outperforms mainstream baseline methods in training efficiency.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

Federated Learning (FL) enables participant devices to collaboratively train deep learning models without sharing their data with the server or other devices, effectively addressing data privacy and computational concerns. However, FL faces a major bottleneck due to high communication overhead from frequent model updates between devices and the server, limiting deployment in resource-constrained wireless networks. In this paper, we propose a three-fold strategy. Firstly, an Adaptive Feature-Elimination Strategy to drop less important features while retaining high-value ones; secondly, Adaptive Gradient Innovation and Error Sensitivity-Based Quantization, which dynamically adjusts the quantization level for innovative gradient compression; and thirdly, Communication Frequency Optimization to enhance communication efficiency. We evaluated our proposed model's performance through extensive experiments, assessing accuracy, loss, and convergence compared to baseline techniques. The results show that our model achieves high communication efficiency in the framework while maintaining accuracy.

Problem

• Reducing communication overhead in Federated Learning systems
• Optimizing gradient quantization for efficient model updates
• Balancing communication frequency with model accuracy preservation

Innovation

• Adaptive feature elimination for important feature retention
• Dynamic gradient quantization for innovative compression
• Communication frequency optimization enhances efficiency