Optimality of Quantum Randomness Certification with Independent Devices

Xin-Wei Fei,Zhen-Qiang Yin,Chao-Han Cui,Wei Huang,Bing-Jie Xu,Shuang Wang,Wei Chen,Yun-Guang Han,Guang-Can Guo,Zheng-Fu Han
DOI: https://doi.org/10.1364/josab.35.002186
2018-01-01
Journal of the Optical Society of America B
Abstract:Quantum random number generators (QRNGs) play an increasingly important role in cryptography systems. Specifically, QRNG performed in prepare-and-measure scenarios with independent devices has attracted intensive study, since it can be intrinsically loss tolerant and its assumptions on devices seem to be quite natural. However, the estimations of min-entropy given in previous works are not optimal. As a result, the performance of quantum randomness certification in practice is still severely depressed. Here, we present a novel method to estimate the min-entropy through developing a reduced device model and proving its equivalence to the original one. Benefitting from this reduction, we can solve the min-entropy estimation problem using mixed integer programming effectively. Furthermore, we derive an optimal bound of analytic form in the case that constraints are made up of symmetric observed probabilities. Simulation results show that our work gives an optimal solution in this case and brings a significant performance improvement compared to previous works.
What problem does this paper attempt to address?