Model Partition and Resource Allocation for Split Learning in Vehicular Edge Networks

TL;DR

This paper introduces a U-shaped split federated learning framework for vehicular edge networks that enhances privacy, reduces communication costs, and optimizes resource allocation using deep reinforcement learning.

Contribution

It proposes a novel U-SFL framework with semantic-aware auto-encoder and DRL-based resource optimization tailored for vehicular edge networks.

Findings

U-SFL achieves similar accuracy to traditional split learning.

Significant reduction in data transmission volume and latency.

Effective DRL algorithm balances latency, energy, and performance.

Abstract

The integration of autonomous driving technologies with vehicular networks presents significant challenges in privacy preservation, communication efficiency, and resource allocation. This paper proposes a novel U-shaped split federated learning (U-SFL) framework to address these challenges on the way of realizing in vehicular edge networks. U-SFL is able to enhance privacy protection by keeping both raw data and labels on the vehicular user (VU) side while enabling parallel processing across multiple vehicles. To optimize communication efficiency, we introduce a semantic-aware auto-encoder (SAE) that significantly reduces the dimensionality of transmitted data while preserving essential semantic information. Furthermore, we develop a deep reinforcement learning (DRL) based algorithm to solve the NP-hard problem of dynamic resource allocation and split point selection. Our comprehensive evaluation demonstrates that U-SFL achieves comparable classification performance to traditional split learning (SL) while substantially reducing data transmission volume and communication latency. The proposed DRL-based optimization algorithm shows good convergence in balancing latency, energy consumption, and learning performance.