差分可辨性隐私参数的迭代分配方法

在大数据时代,对个人隐私的保护不容忽视.于2013年被提出的ρ-差分可辨性定义解决了传统差分隐私仅关注个体对数据库输出影响的问题,使隐私保护的重点转移到防止个体被重新识别上,更加符合相关法律的定义.然而,现阶段对差分可辨性的相关研究较少.本文基于差分可辨性组合性质,提出了差分可辨性隐私参数的迭代分配方法,能够在迭代轮数固定和未知两种情况下分配差分可辨性隐私保护参数,使最终模型满足差分可辨性的隐私定义.对于某些需要迭代的模型,如聚类算法k-means,在聚类过程中可能出现隐私泄露,可以借助差分可辨性的实现机制来对每轮迭代进行加噪处理来保护隐私.实验结果表明,本文方法对数据进行噪声添加后,一定程度上能够保证经过差分可辨性隐私保护的聚类结果可用性.

No-relationship Between Impossibility of Faster-Than-Light Quantum Communication and Distinction of Ensembles with the Same Density Matrix

It has been claimed in the literature that impossibility of faster-than-light quantum communication has an origin of indistinguishability of ensembles with the same density matrix. We show that the two concepts are not related.We argue that even with an ideal single-atom-precision measurement, it is generally impossible to produce two ensembles with exactly the same density matrix; or to produce ensembles with the same density matrix, classical communication is necessary. Hence the impossibility of faster-than-light communication does not imply the indistinguishability of ensembles with the same density matrix.