CDQKL: Consensus-Based Distributed Quantum Kernel Learning

The field of quantum computing has developed rapidly in recent years due to its promising trend of surpassing traditional machine learning in terms of speed and effectiveness. Quantum kernel learning is one of the paradigms of quantum machine learning, but the training of quantum kernel is time consuming. Therefore, this work makes the first attempt to introduce a consensus-based distributed approach to quantum kernel learning - named CDQKL - that only requires to exchange model parameter information between adjacent nodes while avoiding the need of sharing local training data. Through comparative experimental studies, the advantages of CDQKL in classification accuracy and convergence speed are verified. Considering the popularization of quantum computing cloud service and miniaturization of quantum terminals, the CDQKL adapting to this trend is able to play a vital role in data security, which implies the far-reaching significance of this work. Our code is available at https://github.com/Leisurivan/CDOKL.