Visible Rank and Codes with Locality
Abstract
We propose a framework to study the effect of local recovery requirements of codeword symbols on the dimension of linear codes, based on a combinatorial proxy that we call \emph{visible rank}. The locality constraints of a linear code are stipulated by a matrix of 's and 's (which we call a "stencil"), whose rows correspond to the local parity checks (with the 's indicating the support of the check). The visible rank of is the largest for which there is a submatrix in with a unique generalized diagonal of 's. The visible rank yields a field-independent combinatorial lower bound on the rank of and thus the co-dimension of the code. We prove a rank-nullity type theorem relating visible rank to the rank of an associated construct called \emph{symmetric spanoid}, which was introduced by Dvir, Gopi, Gu, and Wigderson~\cite{DGGW20}. Using this connection and a construction of appropriate stencils, we answer a question posed in \cite{DGGW20} and demonstrate that symmetric spanoid rank cannot improve the currently best known upper bound on the dimension of -query locally correctable codes (LCCs) of length . We also study the -Disjoint Repair Group Property (-DRGP) of codes where each codeword symbol must belong to disjoint check equations. It is known that linear -DRGP codes must have co-dimension . We show that there are stencils corresponding to -DRGP with visible rank as small as . However, we show the second tensor of any -DRGP stencil has visible rank , thus recovering the lower bound for -DRGP. For -LCC, however, the 'th tensor power for is unable to improve the upper bound on the dimension of -LCCs by a polynomial factor.
Cite
@article{arxiv.2108.12687,
title = {Visible Rank and Codes with Locality},
author = {Omar Alrabiah and Venkatesan Guruswami},
journal= {arXiv preprint arXiv:2108.12687},
year = {2022}
}
Comments
22 pages; Appeared in RANDOM'21; The current version includes Theorem 5, which is a solution to Question 2 that was asked in the earlier version