English

Matrix discrepancy and the log-rank conjecture

Combinatorics 2023-12-01 v1 Computational Complexity

Abstract

Given an m×nm\times n binary matrix MM with M=pmn|M|=p\cdot mn (where M|M| denotes the number of 1 entries), define the discrepancy of MM as \mboxdisc(M)=maxX[m],Y[n]M[X×Y]pXY\mbox{disc}(M)=\displaystyle\max_{X\subset [m], Y\subset [n]}\big||M[X\times Y]|-p|X|\cdot |Y|\big|. Using semidefinite programming and spectral techniques, we prove that if \mboxrank(M)r\mbox{rank}(M)\leq r and p1/2p\leq 1/2, then \mboxdisc(M)Ω(mn)min{p,p1/2r}.\mbox{disc}(M)\geq \Omega(mn)\cdot \min\left\{p,\frac{p^{1/2}}{\sqrt{r}}\right\}. We use this result to obtain a modest improvement of Lovett's best known upper bound on the log-rank conjecture. We prove that any m×nm\times n binary matrix MM of rank at most rr contains an (m2O(r))×(n2O(r))(m\cdot 2^{-O(\sqrt{r})})\times (n\cdot 2^{-O(\sqrt{r})}) sized all-1 or all-0 submatrix, which implies that the deterministic communication complexity of any Boolean function of rank rr is at most O(r)O(\sqrt{r}).

Keywords

Cite

@article{arxiv.2311.18524,
  title  = {Matrix discrepancy and the log-rank conjecture},
  author = {Benny Sudakov and István Tomon},
  journal= {arXiv preprint arXiv:2311.18524},
  year   = {2023}
}

Comments

9 pages

R2 v1 2026-06-28T13:36:54.749Z