Machine learning continuous-variable quantum key distribution embedded with quantum multi-label classifier. (arXiv:1906.03779v2 [quant-ph] UPDATED)

We propose a brand-new protocol called machine learning continuous-variable
quantum key distribution (ML-CVQKD), aiming to break the limitation of
traditional pattern in CVQKD and establish a cross research platform between
CVQKD and machine learning. ML-CVQKD divides the whole system into state
learning process and state prediction process. The former is used for training
and estimating quantum classifier, and the latter is used for predicting
unlabeled signal states. Meanwhile, a quantum multi-label classification (QMLC)
algorithm is designed as an embedded classifier for ML-CVQKD. Feature
extraction for coherent state and machine learning-based criteria for CVQKD are
successively suggested. QMLC-embedded ML-CVQKD protocol could improve the
performance of CVQKD system in terms of both secret key rate and transmission
distance, and it provides a novel approach for improving the practical CVQKD
system.

Article web page: