Lecture 2- assumptions regarding A'

Lecture 2- assumptions regarding A'

Hila Chefer 09 Nov 2019 22:04

When we proved that the new scheme with E' was a computationally secure, we assumed towards contradiction that there existed an efficient attacker for E' who can tell encryptions apart with polynomial probability. Why did we have to assume it for infinitely many n? Doesn't a single n suffice to contradict the definition?

Thanks!

Re: Lecture 2- assumptions regarding A'

nbitansky 10 Nov 2019 07:37

Formally, to contradict security, you would have to show that there does not exist a negligible $mu$ such that the chance of breaking is always bounded by $mu$(n)]]. Negligible is an asymptotic term, it says that [[$\mu$ will eventually be smaller than any polynomial. (In fact, showing that a breaker works with some polynomial probability for a specific $n$ is in fact vacuous, try to think why.)