个性化联邦蒸馏辅助的车辆边缘缓存策略
机器学习
2026-04-07 v2
摘要
车辆边缘缓存是一项有前景的技术，可通过将用户感兴趣的内容预先缓存在边缘节点来显著降低车辆用户访问内容的延迟。在不暴露用户隐私的前提下准确预测车辆用户感兴趣的内容至关重要。传统的联邦学习通过共享模型而非原始数据来保护用户隐私。然而，联邦学习的训练需要频繁的模型传输，可能导致显著的通信开销。此外，车辆可能在训练完成前离开路侧单元覆盖范围，导致训练失败。为解决这些问题，本文提出了一种个性化联邦蒸馏辅助的车辆边缘缓存策略。仿真结果表明，所提出的车辆边缘缓存策略对车速变化具有良好的鲁棒性，能显著降低通信开销。
引用
@article{arxiv.2512.09378,
title = {Personalized Federated Distillation Assisted Vehicle Edge Caching Strategy},
author = {Xun Li and Qiong Wu and Pingyi Fan and Kezhi Wang and Wen Chen and Cui Zhang},
journal= {arXiv preprint arXiv:2512.09378},
year = {2026}
}
备注
This paper has been accepted by IEEE International Conference on Radio Frequency and Antenna Technologies. The source code has been released at: https://github.com/qiongwu86/Federated-Distillation-Assisted-Vehicle-Edge-Caching-Scheme-Based-on-Lightweight-DDPM