arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1905.11963cs.LGcs.DMstat.ML
收藏

动态图回归的次线性更新时间随机算法

Sublinear Update Time Randomized Algorithms for Dynamic Graph Regression

Amirkabir University of Technology (Tehran Polytechnic)(阿米尔卡比尔理工大学(德黑兰理工学院))

Mostafa Haghir Chehreghani

摘要:

数据科学和机器学习中的一个众所周知的问题是{ \em线性回归},它最近被扩展到动态图。现有的精确算法更新动态图回归的解决方案需要至少一个线性的时间($n$:图的大小)。然而,这种时间复杂度在实践中可能是难以处理的。本文利用{ \em下采样随机Hadamard变换}和\textsf{CountSketch}提出了一般动态图回归的第一个次线性更新时间随机算法。假设给定一个$n\times d$矩阵嵌入图的$\mathbf M$,其中$d \ll n$和$\mathbf M$具有一定的性质.令$r$是用于$1\pm ε$近似的子采样随机化Hadamard变换所需的样本数,其是$n$的次线性。我们的第一个算法支持边插入和边删除,并在$O(rd)$时间更新近似解。第二种算法基于\textsf{CountSketch},支持边插入、边删除、节点插入和节点删除。它以$O(qd)$时间更新近似解,其中$q=O\left(\frac{d^2}{ε^2} \log^6(d/ε) \right)$。

英文摘要:

A well-known problem in data science and machine learning is {\em linear regression}, which is recently extended to dynamic graphs. Existing exact algorithms for updating the solution of dynamic graph regression require at least a linear time (in terms of $n$: the size of the graph). However, this time complexity might be intractable in practice. In the current paper, we utilize {\em subsampled randomized Hadamard transform} and \textsf{CountSketch} to propose the first sublinear update time randomized algorithms for regression of general dynamic graphs. Suppose that we are given a $n\times d$ matrix embedding $\mathbf M$ of the graph, where $d \ll n$ and $\mathbf M$ has certain properties. Let $r$ be the number of samples required by subsampled randomized Hadamard transform for a $1\pm ε$ approximation, which is a sublinear of $n$. Our first algorithm supports edge insertion and edge deletion and updates the approximate solution in $O(rd)$ time. Our second algorithm is based on \textsf{CountSketch} and supports edge insertion, edge deletion, node insertion and node deletion. It updates the approximate solution in $O(qd)$ time, where $q=O\left(\frac{d^2}{ε^2} \log^6(d/ε) \right)$.

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。

↑