SL-FAC: A Communication-Efficient Split Learning Framework with Frequency-Aware Compression

2026-04-08

AI Summary

This work addresses the high communication overhead in split learning caused by frequent transmission of activations and gradients. To mitigate this issue, the authors propose a frequency-domain adaptive compression framework that transforms intermediate data into the frequency domain and performs adaptive frequency decomposition (AFD) based on spectral energy distribution. A frequency-aware quantization strategy (FQC) is further introduced to apply differentiated compression across distinct frequency bands. This approach substantially reduces communication volume while preserving essential training information, thereby achieving a joint optimization of communication efficiency and model convergence performance. Experimental results demonstrate that the proposed framework significantly lowers communication costs in multi-device, high-complexity scenarios without compromising model accuracy or convergence speed.

Abstract

The growing complexity of neural networks hinders the deployment of distributed machine learning on resource-constrained devices. Split learning (SL) offers a promising solution by partitioning the large model and offloading the primary training workload from edge devices to an edge server. However, the increasing number of participating devices and model complexity leads to significant communication overhead from the transmission of smashed data (e.g., activations and gradients), which constitutes a critical bottleneck for SL. To tackle this challenge, we propose SL-FAC, a communication-efficient SL framework comprising two key components: adaptive frequency decomposition (AFD) and frequency-based quantization compression (FQC). AFD first transforms the smashed data into the frequency domain and decomposes it into spectral components with distinct information. FQC then applies customized quantization bit widths to each component based on its spectral energy distribution. This collaborative approach enables SL-FAC to achieve significant communication reduction while strategically preserving the information most crucial for model convergence. Extensive experiments confirm the superior performance of SL-FAC for improving the training efficiency.

Problem

• split learning
• communication overhead
• smashed data
• resource-constrained devices
• distributed machine learning

Innovation

• Split Learning
• Frequency-Aware Compression
• Adaptive Frequency Decomposition
• Quantization
• Communication Efficiency