Non-asymptotic analysis of privacy amplification via Rényi entropy and inf-spectral entropy
Non-asymptotic analysis of privacy amplification via Rényi entropy and inf-spectral entropy
复制标题
DOI：
10.1109/isit.2013.6620720
复制
发表时间：
2012-11
期刊：
影响因子：
--
通讯作者：
Shun Watanabe;Masahito Hayashi
中科院分区：
文献类型：
--
作者：
Shun Watanabe;Masahito Hayashi
本文研究了隐私放大问题，并比较了现有的两个界：由作者之一导出的指数界和Renner导出的最小熵界。事实证明，指数界是优于最小熵界时，安全参数是相当小的块长度，和最小熵界是优于指数界时，安全参数是相当大的块长度。此外，我们提出了另一个界，插值的指数界和最小熵界的混合使用的雷诺熵和inf-spectral熵。
This paper investigates the privacy amplification problem, and compares the existing two bounds: the exponential bound derived by one of the authors and the min-entropy bound derived by Renner. It turns out that the exponential bound is better than the min-entropy bound when a security parameter is rather small for a block length, and that the min-entropy bound is better than the exponential bound when a security parameter is rather large for a block length. Furthermore, we present another bound that interpolates the exponential bound and the min-entropy bound by a hybrid use of the Rényi entropy and the inf-spectral entropy.