English

Efficient algorithms for certifying lower bounds on the discrepancy of random matrices

Data Structures and Algorithms 2023-06-02 v2 Computational Complexity Discrete Mathematics

Abstract

We initiate the study of the algorithmic problem of certifying lower bounds on the discrepancy of random matrices: given an input matrix ARm×nA \in \mathbb{R}^{m \times n}, output a value that is a lower bound on disc(A)=minx{±1}nAx\mathsf{disc}(A) = \min_{x \in \{\pm 1\}^n} ||Ax||_\infty for every AA, but is close to the typical value of disc(A)\mathsf{disc}(A) with high probability over the choice of a random AA. This problem is important because of its connections to conjecturally-hard average-case problems such as negatively-spiked PCA, the number-balancing problem and refuting random constraint satisfaction problems. We give the first polynomial-time algorithms with non-trivial guarantees for two main settings. First, when the entries of AA are i.i.d. standard Gaussians, it is known that disc(A)=Θ(n2n/m)\mathsf{disc} (A) = \Theta (\sqrt{n}2^{-n/m}) with high probability. Our algorithm certifies that disc(A)exp(O(n2/m))\mathsf{disc}(A) \geq \exp(- O(n^2/m)) with high probability. As an application, this formally refutes a conjecture of Bandeira, Kunisky, and Wein on the computational hardness of the detection problem in the negatively-spiked Wishart model. Second, we consider the integer partitioning problem: given nn uniformly random bb-bit integers a1,,ana_1, \ldots, a_n, certify the non-existence of a perfect partition, i.e. certify that disc(A)1\mathsf{disc} (A) \geq 1 for A=(a1,,an)A = (a_1, \ldots, a_n). Under the scaling b=αnb = \alpha n, it is known that the probability of the existence of a perfect partition undergoes a phase transition from 1 to 0 at α=1\alpha = 1; our algorithm certifies the non-existence of perfect partitions for some α=O(n)\alpha = O(n). We also give efficient non-deterministic algorithms with significantly improved guarantees. Our algorithms involve a reduction to the Shortest Vector Problem.

Keywords

Cite

@article{arxiv.2211.07503,
  title  = {Efficient algorithms for certifying lower bounds on the discrepancy of random matrices},
  author = {Prayaag Venkat},
  journal= {arXiv preprint arXiv:2211.07503},
  year   = {2023}
}

Comments

ITCS 2023

R2 v1 2026-06-28T05:49:23.950Z