LSV-Based Tail Inequalities for Sums of Random Matrices

LSV-Based Tail Inequalities for Sums of Random Matrices
复制标题

基于 LSV 的随机矩阵和的尾部不等式

DOI:
10.1162/neco_a_00901
复制
发表时间:
2017
期刊:
影响因子:
2.9
通讯作者:
Dacheng Tao
Dacheng Tao
中科院分区:
计算机科学4区
文献类型:
--
作者:
Chao Zhang;杜磊;Dacheng Tao

文献摘要

参考文献

相似文献

随机矩阵技术在许多机器学习模型中发挥了重要作用。在这封信中,我们提出了一种研究随机矩阵和的尾不等式的新方法。与其他工作(AhlSwede&温特,2002;Tropp,2012;Hsu,Kakade,&Zhang,2012)不同,我们的尾部结果是基于最大奇异值(LSV)的,并且与矩阵的维度无关。由于LSV运算和期望是非对易的,我们引入了一种对角化方法,将LSV运算转化为无穷维对角矩阵的迹运算。通过这种方法,我们得到了另一种形式的拉普拉斯变换界,进而得到了随机矩阵和的基于LSV的尾部不等式。
The techniques of random matrices have played an important role in many machine learning models. In this letter, we present a new method to study the tail inequalities for sums of random matrices. Different from other work (Ahlswede & Winter, 2002; Tropp, 2012; Hsu, Kakade, & Zhang, 2012), our tail results are based on the largest singular value (LSV) and independent of the matrix dimension. Since the LSV operation and the expectation are noncommutative, we introduce a diagonalization method to convert the LSV operation into the trace operation of an infinitely dimensional diagonal matrix. In this way, we obtain another version of Laplace-transform bounds and then achieve the LSV-based tail inequalities for sums of random matrices.
DOI: 10.1214/13-aop892
发表时间: 2012-01
影响因子: 2.3
作者:
Lester W. Mackey;Michael I. Jordan;Richard Y. Chen;Brendan Farrell;J. Tropp
通讯作者: Lester W. Mackey;Michael I. Jordan;Richard Y. Chen;Brendan Farrell;J. Tropp
DOI: 10.1002/0470871296
发表时间: 2004-01
影响因子: 11.1
作者:
R. S. Cantrell;C. Cosner
通讯作者: R. S. Cantrell;C. Cosner
DOI: 10.1214/ecp.v17-1869
发表时间: 2012-12
影响因子: 0.5
作者:
Daniel J. Hsu;S. Kakade;Tong Zhang
通讯作者: Daniel J. Hsu;S. Kakade;Tong Zhang
DOI: 10.1103/physrevlett.97.188104
发表时间: 2006-11-03
影响因子: 8.6
作者:
Rajan, Kanaka;Abbott, L. F.
通讯作者: Abbott, L. F.
DOI: --
发表时间: 2003-11
期刊: --
影响因子: --
作者:
R. S. Cantrell;C. Cosner
通讯作者: R. S. Cantrell;C. Cosner