QFMLNet: An AI-Driven Quantum Communication Architecture for 6G Space-Ground Networks

Xufei Ding, Yingchi Mao, Jianan Huang, Shuhua Huang, Zicheng Wang, Jie Wu

Abstract: Quantum communication in 6G offers unprecedented data transmission speeds, ultra-low latency, and seamless space-ground connectivity. AI and quantum communication technologies, including quantum machine learning (QML) and quantum federated learning (QFL), play a key role in enhancing secure communication and data processing despite challenges like high resource demands and unique security risks. To overcome these challenges, we propose QFMLNet, a dedicated framework that enables a decentralized quantum neural network (QNN) training on protected data. QFMLNet is a secure and communication-efficient framework that combines quantum homomorphic encryption and quantum key distribution as its security mechanism. It enables the edge to train its private QNN model and transmit it to the server for aggregation through quantum homomorphic encryption. Each framework layer is designed to address specific needs in 6G quantum networks, enhancing model robustness, safeguarding communication, and ensuring reliable performance in complex, security-sensitive environments. The framework enhances model robustness and safeguards communication, offering a structured solution for secure quantum communication in 6G space-ground networks.