English

Skew-sparse matrix multiplication

Computational Complexity 2022-05-16 v1 Numerical Analysis Symbolic Computation Numerical Analysis

Abstract

Based on the observation that Q(p1)×(p1)\mathbb{Q}^{(p-1) \times (p-1)} is isomorphic to a quotient skew polynomial ring, we propose a new method for (p1)×(p1)(p-1)\times (p-1) matrix multiplication over Q\mathbb{Q}, where pp is a prime number. The main feature of our method is the acceleration for matrix multiplication if the product is skew-sparse. Based on the new method, we design a deterministic algorithm with complexity O(Tω2p2)O(T^{\omega-2} p^2), where Tp1T\le p-1 is a parameter determined by the skew-sparsity of input matrices and ω\omega is the asymptotic exponent of matrix multiplication. Moreover, by introducing randomness, we also propose a probabilistic algorithm with complexity O(tω2p2+p2log1ν)O^\thicksim(t^{\omega-2}p^2+p^2\log\frac{1}{\nu}), where tp1t\le p-1 is the skew-sparsity of the product and ν\nu is the probability parameter.

Keywords

Cite

@article{arxiv.2205.06429,
  title  = {Skew-sparse matrix multiplication},
  author = {Qiao-Long Huang and Ke Ye and Xiao-Shan Gao},
  journal= {arXiv preprint arXiv:2205.06429},
  year   = {2022}
}
R2 v1 2026-06-24T11:16:07.851Z