More Efficient Privacy Amplification with Non-Uniform Random Seeds via Dual Universal Hash Function

We explicitly construct random hash functions for privacy amplification that requires random seeds below the previous theoretical bound and allows efficient implementations. When the random seeds are not the uniform random numbers, we evaluate the security parameter of the final keys with the above random hash function as well as the random hash function constructed by the modified Toeplitz matrix by using the minimum entropy of the random seeds. The key idea is the concept of dual universal_2 hash function, introduced in our previous paper.