Optimal resource allocation of quantum digital signatures with machine learning

Quantum digital signature is one of the most promising quantum technologies, which can guarantee the authenticity and transferability of messages based on the law of quantum mechanics. Up to date, various protocols have been proposed to make quantum digital signature more practical. However, most of them consider only single protocol and security, neglecting resource allocations in practical applications, especially under the environment of multi-party real-time signatures. Here, we for the first time implement the machine learning method based on the random forest algorithm into the optimization of quantum digital signature systems, and with that, we can predict not only the most suitable protocol, but also the optimal system parameters in real time, with an accuracy of more than 97%.