arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.05758math.NAcs.NA
收藏

基于稀疏图的快速数值线性代数绘制

Sparse graph based sketching for fast numerical linear algebra

Dong Hu, Shashanka Ubaru, Alex Gittens, Kenneth L. Clarkson, Lior Horesh, Vassilis Kalantzis

摘要:

近年来,各种各样的随机结构的草图矩阵已经被设计出来,已被用于快速算法的数值线性代数的问题,如最小二乘回归,低秩近似,和近似的杠杆得分。草图矩阵的一个关键属性是子空间嵌入。在本文中,我们研究了从稀疏的二分图中获得的草图矩阵,即,剩下的度数很小。特别是,我们探讨了两个流行的类稀疏图,即扩展图和神奇的图。对于给定的子空间$\mathcal{U} \subseteq \mathbb{R}^n$的尺寸$k$,我们表明,神奇的图与左度$s=2$产生$(1\pm ε)$ ${\ell}_2$-子空间嵌入$\mathcal{U}$,如果右顶点的数量(草图大小)$m = \mathcal{O}({k^2}/{ε^2})$。具有$s = \mathcal{O}({\log k}/ε)$的扩展图产生$m = \mathcal{O}({k \log k}/{ε^2})$的子空间嵌入。我们还讨论了稀疏素描矩阵的建设与减少随机性使用扩展器的基础上纠错码。在各种合成数据集和真实数据集上的实验结果表明,这些稀疏图草图矩阵在实践中工作得很好。

英文摘要:

In recent years, a variety of randomized constructions of sketching matrices have been devised, that have been used in fast algorithms for numerical linear algebra problems, such as least squares regression, low-rank approximation, and the approximation of leverage scores. A key property of sketching matrices is that of subspace embedding. In this paper, we study sketching matrices that are obtained from bipartite graphs that are sparse, i.e., have left degree~s that is small. In particular, we explore two popular classes of sparse graphs, namely, expander graphs and magical graphs. For a given subspace $\mathcal{U} \subseteq \mathbb{R}^n$ of dimension $k$, we show that the magical graph with left degree $s=2$ yields a $(1\pm ε)$ ${\ell}_2$-subspace embedding for $\mathcal{U}$, if the number of right vertices (the sketch size) $m = \mathcal{O}({k^2}/{ε^2})$. The expander graph with $s = \mathcal{O}({\log k}/ε)$ yields a subspace embedding for $m = \mathcal{O}({k \log k}/{ε^2})$. We also discuss the construction of sparse sketching matrices with reduced randomness using expanders based on error-correcting codes. Empirical results on various synthetic and real datasets show that these sparse graph sketching matrices work very well in practice.

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。

↑