Source-linked AI summary

Practical and fast quantum random number generation based on photon arrival time relative to external reference

You-Qi Nie, Hong-Fei Zhang, Zhen Zhang, Jian Wang, Xiongfeng Ma, Jun Zhang, Jian-Wei Pan

arXiv:1401.2594v1quant-phcs.CR

TL;DR

Successive-detection interval schemes can produce biased raw bits, while reducing generation rate; hardware shaping adds system complexity. The scheme records photon detection time relative to an external reference, models min-entropy while accounting for device imperfections, and applies Toeplitz hashing for extraction. The raw rate reached 109 Mbps and extracted rate 96 Mbps; over 1 Tb was acquired, with 30 randomly selected 1 Gb files and other tested files passing NIST tests.

  • Problem

    Successive-detection interval schemes can produce biased raw bits, while reducing generation rate; hardware shaping adds system complexity.

    Source: p. 2
  • Method

    The scheme records photon detection time relative to an external reference, models min-entropy while accounting for device imperfections, and applies Toeplitz hashing for extraction.

    Source: pp. 3, 4, 6
  • Results

    The raw rate reached 109 Mbps and extracted rate 96 Mbps; over 1 Tb was acquired, with 30 randomly selected 1 Gb files and other tested files passing NIST tests.

    Source: pp. 6–7
  • Takeaways & Limitations

    The authors conclude that the scheme's simplicity and robustness suit practical applications.

    Source: p. 8

    Multi-photon emission and detector imperfections can lower raw-data min-entropy, and the extraction accounts for these effects.

    Source: pp. 4–5

Abstract

from arXiv · show

We present a practical high-speed quantum random number generator, where the timing of single-photon detection relative to an external time reference is measured as the raw data. The bias of the raw data can be substantially reduced compared with the previous realizations. The raw random bit rate of our generator can reach 109 Mbps. We develop a model for the generator and evaluate the min-entropy of the raw data. Toeplitz matrix hashing is applied for randomness extraction, after which the final random bits are able to pass the standard randomness tests.

Loading 1401.2594v1…