Beating Fredman-Koml\'{o}s for perfect $k$-hashing
Abstract
We say a subset is a -hash code (also called -separated) if for every subset of codewords from , there exists a coordinate where all these codewords have distinct values. Understanding the largest possible rate (in bits), defined as , of a -hash code is a classical problem. It arises in two equivalent contexts: (i) the smallest size possible for a perfect hash family that maps a universe of elements into , and (ii) the zero-error capacity for decoding with lists of size less than for a certain combinatorial channel. A general upper bound of on the rate of a -hash code (in the limit of large ) was obtained by Fredman and Koml\'{o}s in 1984 for any . While better bounds have been obtained for , their original bound has remained the best known for each . In this work, we obtain the first improvement to the Fredman-Koml\'{o}s bound for every . While we get explicit (numerical) bounds for , for larger we only show that the FK bound can be improved by a positive, but unspecified, amount. Under a conjecture on the optimum value of a certain polynomial optimization problem over the simplex, our methods allow an effective bound to be computed for every .
Keywords
Cite
@article{arxiv.1805.04151,
title = {Beating Fredman-Koml\'{o}s for perfect $k$-hashing},
author = {Venkatesan Guruswami and Andrii Riazanov},
journal= {arXiv preprint arXiv:1805.04151},
year = {2018}
}
Comments
19 pages