Integrating quantum key distribution (QKD) with existing optical networks is highly desired to reduce the deployment costs and achieve efficient resource utilization, and some point-to-point transmitting experiments have verified its feasibility. Nevertheless, there are still many problems in the realistic scenario where QKD coexists with dynamic data traffics. On the one hand, the conventional static channel allocation schemes cannot guarantee the quality of quantum channels in the presence of the time-varying noises. On the other hand, considering the complex noise generation caused by dynamic classical data traffics with variable characters, it is challenging to achieve online high-performance quantum channel assignments. To address these problems, we propose a machine learning based noise-suppressing channel allocation (ML-NSCA) scheme. In this scheme, the LightGBM based ML framework is trained to predict the optimal channel allocations with lowest noise impacts, according to which, the quantum channels are periodically reallocated to guarantee high secure key rate. To improve the accuracy and scalability of the ML framework, we also optimize the method of feature extraction during the training process. The performance evaluation results indicate that the proposed scheme can effectively resist the dynamic noise impacts in the realistic optical networks and obtain higher secure key rate with less operation complexity than the previous schemes.
Integrating quantum key distribution (QKD) with existing optical networks is highly desired to reduce the deployment costs and achieve efficient resource utilization, and some point-to-point transmitting experiments have verified its feasibility. Nevertheless, there are still many problems in the realistic scenario where QKD coexists with dynamic data traffics. On the one hand, the conventional static channel allocation schemes cannot guarantee the quality of quantum channels in the presence of the time-varying noises. On the other hand, considering the complex noise generation caused by dynamic classical data traffics with variable characters, it is challenging to achieve online high-performance quantum channel assignments. To address these problems, we propose a machine learning based noise-suppressing channel allocation (ML-NSCA) scheme. In this scheme, the LightGBM based ML framework is trained to predict the optimal channel allocations with lowest noise impacts, according to which, the quantum channels are periodically reallocated to guarantee high secure key rate. To improve the accuracy and scalability of the ML framework, we also optimize the method of feature extraction during the training process. The performance evaluation results indicate that the proposed scheme can effectively resist the dynamic noise impacts in the realistic optical networks and obtain higher secure key rate with less operation complexity than the previous schemes.
Quantum key distribution (QKD) is a cutting-edge technology that guarantees unbreakable security. Multi-band transmission across the O+E+S+C+L bands offers a viable solution for the coexistence of quantum and classical signals over existing fiber infrastructure. However, the secure key rate (SKR) achievable in quantum channels (QChs) is influenced by variations in classical traffic load and its spectrum usage patterns. To support dynamic and time-varying classical traffic, it is essential to estimate the achievable SKR for each QCh in real-time, enabling the selection of the optimal frequency that maximizes SKR. Conventional methods rely on solving complex integral noise equations to estimate SKR, but their computational complexity makes them unsuitable for real-time operations. In this letter, we propose a machine learning (ML) algorithm to evaluate the SKR of QChs, taking into account the time-varying behavior of classical traffic, and to select the optimal frequency for QChs. We implement three ML algorithms across various fiber intervals, all of which estimate the optimal frequency for QChs with 99% accuracy and perform computations in an average of 0.09 seconds— significantly faster than the conventional method, which has a mean computation time of 637 seconds.
Conducting quantum key distribution (QKD) through existing optical fibers together with conventional communication signals is a viable way to expand its practical application, but weak quantum signals can be severely disrupted by co-propagating classical signals. In this paper, the suppression of four-wave mixing (FWM) noise and Raman noise is considered simultaneously for the first time, to the best of our knowledge, and the joint optimized channel allocation (JOCA) scheme is proposed. In the JOCA scheme, the quantum channels and classical channels are interleaved with each other to avoid FWM noise and optimal quantum channel positions are chosen in variable conditions according to the Raman scattering spectrum. Experimental measurements of the noise photons show that the JOCA scheme can effectively reduce the impairments on quantum signals compared with the single-target schemes. Additionally, simulation results verify that the JOCA scheme can increase the secure key generation rate and transmission distance, and that it also enables the DWDM-QKD system to tolerate higher-power classical signals and more classical channels, which improve the compatibility with a high-capacity communication system.
The integration of Quantum Key Distribution (QKD) with Machine Learning (ML) has developed a high-end, secure model for accurate data transmission and network security. Many existing systems face several issues due to advances in quantum computing. Thus, integrating these models improved the efficiency, accuracy, and feasibility of QKD systems. In this paper, an Advanced Learning Approach (ALA) is combined with the Quantum Key Distribution Framework to refine key generation, error correction, and intrusion detection in quantum communication channels. The proposed ALA also used reinforcement learning algorithms to predict quantum channel fluctuations, dynamically adjust key exchange parameters, and strengthen authentication protocols. Simulation results demonstrate improved key rates, reduced Quantum Bit Error Rates (QBER), and enhanced resilience against eavesdropping attacks. The study contributes to the realization of practical, scalable, and intelligent quantum-secure communication systems—bridging the gap between theoretical quantum security and real-world network deployment.
Quantum Key Distribution (QKD) offers theoretically unbreakable security by leveraging quantum mechanics. However, practical implementation is challenged by environmental vulnerabilities, noise, and hardware imperfections. Recently, Machine Learning (ML) has emerged as a powerful tool to address these limitations and enhance the real-world viability of QKD systems. In this survey, we review ML techniques applied to improve QKD security and performance across five applications. First, parameter optimization, covering signal calibration, polarization alignment, phase stabilization, modulation state tuning, and post-processing enhancements to maximize secure key generation and minimize error rates. Second, attack detection, where ML models identify and classify quantum threats such as photon-number-splitting and Trojan-horse attacks. Third, protocol selection, leveraging ML to dynamically choose QKD protocols based on operational conditions. Fourth, key performance prediction of core metrics such as Secret Key Rate (SKR) and Quantum Bit Error Rate (QBER). Finally, quantum network management, optimizing large-scale QKD deployments through intelligent routing, node management, and resource allocation. Performance improvements are evaluated using accuracy, reduced QBER, and increased SKR. While ML shows significant potential for finance, government, and defense applications, challenges remain in scalability, computational demands, and real-world testing. Ongoing work should focus on lightweight, generalizable models and standardized benchmarks for practical ML-enhanced QKD deployment.
The fronthaul network is an essential way to improve the comprehensive performance of beyond fifth generation (B5G) communication network. Space-wavelength division multiplexing (SWDM) can effectively improve its capacity, and quantum key distribution technique can provide unconditional information security for it. In quantum secured SWDM B5G fronthaul network, quantum signals are affected by noises generated by classical signals, such as spontaneous Raman scattering, four-wave mixing and inter-core crosstalk. Evaluating these noises in real-time will increase the time delay of the whole network. In this paper, we propose two machine learning (ML) models (XGBoost and LightGBM) to predict these noises. Simulation results show that the ML models can reduce the noise evaluation time by up to 98.8%. Besides, the available channel predicting accuracy rate is close to 100%. The minimum noise predicting accuracy rate increases with the increasement of the channel occupancy rate and can reach 100% when channel occupancy is higher than 80%.