Mutual synchronization of tree parity machines has been proposed as an efficient alternative to key reconciliation in quantum key distribution. However, use of non-binary input vectors, while improving synchronization speed, introduces extrema-value effect, leading to weight clustering at boundary values and reduced key entropy. This paper evaluates joint scaling of synaptic depth L and input range M in tree parity machines with non-binary inputs using Monte Carlo simulations. Performance is assessed in terms of synchronization steps, attacker weight overlap, attacker success probability, boundary weight distribution. The results show that jointly increasing L and M can mitigate the extrema-value effect, reducing the boundary weight percentage by up to 39.3%, and attacker overlap by up to 75%, albeit at the cost of an increased number of synchronization steps. By selecting, for each M group, the L-M configuration with minimum attacker overlap, the study identifies parameter sets that achieve stronger security than the baseline, and example operating points that preserve near-baseline synchronization speed while reducing attacker overlap. These findings indicate that joint tuning of L and M parameters can decrease the extrema-value effect and enhance robustness of tree parity machine-based QKD reconciliation with non-binary inputs, while allowing fast synchronization for suitable parameter choices.
Mutual synchronization of tree parity machines has been proposed as an efficient alternative to key reconciliation in quantum key distribution. However, use of non-binary input vectors, while improving synchronization speed, introduces extrema-value effect, leading to weight clustering at boundary values and reduced key entropy. This paper evaluates joint scaling of synaptic depth L and input range M in tree parity machines with non-binary inputs using Monte Carlo simulations. Performance is assessed in terms of synchronization steps, attacker weight overlap, attacker success probability, boundary weight distribution. The results show that jointly increasing L and M can mitigate the extrema-value effect, reducing the boundary weight percentage by up to 39.3%, and attacker overlap by up to 75%, albeit at the cost of an increased number of synchronization steps. By selecting, for each M group, the L-M configuration with minimum attacker overlap, the study identifies parameter sets that achieve stronger security than the baseline, and example operating points that preserve near-baseline synchronization speed while reducing attacker overlap. These findings indicate that joint tuning of L and M parameters can decrease the extrema-value effect and enhance robustness of tree parity machine-based QKD reconciliation with non-binary inputs, while allowing fast synchronization for suitable parameter choices.
This paper presents a novel approach to enhancing the security of error correction in quantum key distribution by introducing randomization into the update rule of Tree Parity Machines. Two dynamic update algorithms—dynamic_rows and dynamic_matrix—are proposed and tested. These algorithms select the update rule quasi-randomly based on the input vector, reducing the effectiveness of synchronization-based attacks. A series of simulations were conducted to evaluate the security implications under various configurations, including different values of K, N, and L parameters of neural networks. The results demonstrate that the proposed dynamic algorithms can significantly reduce the attacker’s synchronization success rate without requiring additional communication overhead. Both proposed solutions outperformed hebbian, an update rule-based synchronization method utilizing the percentage of attackers synchronization. It has also been shown that when the attacker chooses their update rule randomly, the dynamic approaches work better compared to random walk rule-based synchronization, and that in most cases it is more profitable to use dynamic update rules when an attacker is using random walk. This study contributes to improving QKD’s robustness by introducing adaptive neural-based error correction mechanisms.
Key agreement plays a crucial role in ensuring secure communication in public networks. Although algorithms developed many years ago are still being used, the emergence of quantum computing has prompted the search for new solutions. Tree parity machines have been put forward as a potential solution. However, they possess inherent shortcomings, one of which is the uneven distribution of values in the secured key obtained after the key agreement process, especially when non-binary vectors are used during the synchronization process. This paper introduces a novel algorithm designed to address this issue. The results demonstrate a substantial enhancement in the quality of the secured key obtained.
The article discusses a key reconciliation protocol for quantum key distribution (QKD) systems based on Tree Parity Machines (TPM). The idea of transforming key material into neural network weights is presented. Two experiments were conducted to study how the number of synchronization iterations and the amount of leaked information depend on the quantum bit error rate (QBER) and the range of neural network weights. The results show a direct relationship between the average number of synchronization iterations and QBER, an increase in iterations when the weight range is expanded, and a reduction in leaked information as the weight range increases. Based on these results, conclusions are drawn regarding the applicability of the protocol and the prospects for further research on neural cryptographic methods in the context of key reconciliation.
Key agreement protocols based on neural synchronization with Tree Parity Machines (TPMs) offer promising security advantages: they do not rely on trapdoor functions, making them resistant to quantum attacks, and they avoid the need for specialized hardware required by quantum-based schemes. Nevertheless, these protocols face a significant vulnerability: the large number of public message exchanges required for synchronization increases the risk that an attacker, acting as a Man-in-the-Middle, can successfully synchronize their own TPMs with those of the legitimate parties and ultimately recover the shared key. Motivated by the need to reduce this risk, we propose a novel probabilistic protocol that enables two parties to securely estimate the size of the shared key during intermediate steps, without revealing any key material. This estimation allows the protocol to terminate as soon as sufficient key material has been established, thereby reducing the number of synchronization rounds and limiting the opportunity for an attacker to synchronize. We integrate our estimation mechanism into a neural key agreement protocol and evaluate its performance and security, demonstrating improved efficiency and enhanced resistance to attacks compared to existing approaches.
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.
Quantum key distribution (QKD) is a promising technique for secure communication based on quantum mechanical principles. To improve the secure key rate of a QKD system, most studies on reconciliation primarily focused on improving the efficiency. With the increasing performance of QKD systems, the research priority has shifted to the improvement of both throughput and efficiency. In this paper, we propose a high performance solution of Cascade reconciliation, including a high-throughput-oriented framework and an integrated-optimization-oriented scheme. Benefiting from the fully utilizing computation and storage resources, effectively dealing with communication delays, the integrated-optimization-oriented parameters setting, etc., an excellent overall performance was achieved. Experimental results showed that, the throughput of up to 570Mbps with an efficiency of 1.038 was achieved, which, to our knowledge, was more than four times faster than any throughput previously demonstrated. Furthermore, throughputs on real data sets were capable of reaching up to 86Mbps even on embedded platforms. Additionally, our solution offers good adaptability to the fluctuating communication delay and quantum bit error rate (QBER). Based on our study, low performance (i.e. low power-consumption and cost-effective) CPU platforms will be sufficient for reconciliation in the existing and near-term QKD systems.