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

TL;DR

This paper introduces a semantic-aware auto-encoder (SAE) that significantly reduces the dimensionality of transmitted data while preserving essential semantic information and develops a deep reinforcement learning (DRL) based algorithm to solve the NP-hard problem of dynamic resource allocation and split point selection.

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.

Figures (11)

• Figure 1: System model.
• Figure 2: U-SFL scheme without label sharing, the model is split into three parts (part a, part b and part c) for training.
• Figure 3: Semantic-aware communication framework for U-SFL in vehicular edge networks.
• Figure 4: Model structure and partition points for the deep learning model.
• Figure 5: Training and testing results of different model configurations with different datasets ((a) and (d) for training accuracy, (b) and (e) for testing accuracy). Testing accuracy comparison for different split point combinations ((c) and (f)).