arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2003.03019cs.CCmath.AC
收藏

矩形矩阵乘法的障碍

Barriers for rectangular matrix multiplication

Matthias Christandl, François Le Gall, Vladimir Lysikov, Jeroen Zuiddam

AI总结:

本文针对矩形矩阵乘法,证明现有构造最快算法的方法无法达到复杂度 $n^{p+1}$,并给出精确数值障碍,其中对偶指数 $α$ 的下界不能超过 0.6218。

摘要:

我们研究大规模矩形矩阵相乘的算法问题。我们证明,用于构造矩形矩阵乘法最快算法的方法,无法为 $n \ imes n$ 乘以 $n \ imes n^p$ 的矩阵乘法给出复杂度为 $n^{p + 1}$ 的算法。事实上,我们证明了该方法的一个精确数值障碍。我们的障碍在数值意义上以及在其一般性上都改进了先前已知的障碍。特别地,我们证明,通过大 Coppersmith-Winograd 张量得到的矩阵乘法对偶指数 $α$ 的任何下界都不能超过 0.6218。

英文摘要:

We study the algorithmic problem of multiplying large matrices that are rectangular. We prove that the method that has been used to construct the fastest algorithms for rectangular matrix multiplication cannot give algorithms with complexity $n^{p + 1}$ for $n \times n$ by $n \times n^p$ matrix multiplication. In fact, we prove a precise numerical barrier for this method. Our barrier improves the previously known barriers, both in the numerical sense, as well as in its generality. In particular, we prove that any lower bound on the dual exponent of matrix multiplication $α$ via the big Coppersmith-Winograd tensors cannot exceed 0.6218.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑