Cybersecure Synchronisation of Entangled Quantum Neural Networks With Reconfigurable Intelligent Surface and Quantum Key Distribution for 6G Holographic Communications

Abstract

This paper presents a cybersecure hybrid quantum–classical architecture for synchronising distributed quantum neural networks (QNNs) in 6G holographic communications. The proposed framework targets secure, low-latency and high-fidelity end-to-end operation under realistic noise and adversarial conditions. Multipartite Greenberger–Horne–Zeilinger (GHZ) entanglement supports gradient-consensus synchronisation, regularised by von Neumann entropy and trace distance. Holographic tensor teleportation is executed over reconfigurable intelligent surface (RIS)-assisted midhaul links protected by quantum key distribution (QKD). Joint optimisation of QNN parameters and RIS phases targets high end-to-end fidelity under noise. Qiskit simulations averaged over 100 trials achieve $F_{\text{avg}}=0.961$, $E_{\text{sync}}=0.010$ and $S\left(\rho \right)=0.12$, with 22.5 ms end-to-end latency. Compared with reduced baselines without GHZ synchronisation or without RIS control, fidelity improves by 20%–28% and synchronisation divergence decreases by about 90%. Scalability, security stress scenarios, classical-feedback impairments, hyperparameter sensitivity and noisy intermediate-scale quantum (NISQ)-oriented error mitigation are all evaluated. The architecture scales gracefully to $N=20$ distributed units with only 7% fidelity reduction, whereas mitigation improves fidelity by 7%–11% under moderate noise. Overall, the RIS-QKD-GHZ integration enables secure, low-latency and scalable quantum–classical 6G networking.