中文摘要：

对于语义安全性有两种不同形式的定义,一种基于模拟器,一种基于比较.我们采用Goldreich对语义安全性的基于模拟器的定义,证明了它与相应基于比较的定义都等价于不可区分性,得出了这两种定义确实等价的结论.该结论与Watanabe等人的结论相悖,原因在于他们采用的基于模拟器的语义安全性定义中允许敌手及其模拟器自己选择其输入的有关明文的附加部分信息,导致敌手模拟器与敌手选择的可能不同,这使该定义与不可区分性不等价,与公认的结论不符.

英文摘要：

For semantic security,there are two different definitions.One is simulator-based,the other is comparison-based.For Goldreich＇s simulator-based definition,we show the corresponding comparison-based definition is equivalent to it by proving both of them are equivalent to indistinguishability.This conflicts with that of Watanabe,etc.The reason lies in that in the simulation-based definition they used the adversary and its simulator are allowed to choose the auxiliary partial information about the plaintext to be input by themselves which results in the information chosen by them can be different.This makes the definition even not equivalent to indistinguishability,contradicts with widely accepted result.