Sketching as a Tool for Numerical Linear Algebra

被引:569
作者
Woodruff, David P. [1 ]
机构
[1] IBM Res Almaden, San Jose, CA 95120 USA
来源
FOUNDATIONS AND TRENDS IN THEORETICAL COMPUTER SCIENCE | 2014年 / 10卷 / 1-2期
关键词
D O I
10.1561/0400000060
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
This survey highlights the recent advances in algorithms for numerical linear algebra that have come from the technique of linear sketching, whereby given a matrix, one first compresses it to a much smaller matrix by multiplying it by a (usually) random matrix with certain properties. Much of the expensive computation can then be performed on the smaller matrix, thereby accelerating the solution for the original problem. In this survey we consider least squares as well as robust regression problems, low rank approximation, and graph sparsification. We also discuss a number of variants of these problems. Finally, we discuss the limitations of sketching methods.
引用
收藏
页码:1 / 157
页数:21
相关论文
共 122 条
[1]   Database-friendly random projections: Johnson-Lindenstrauss with binary coins [J].
Achlioptas, D .
JOURNAL OF COMPUTER AND SYSTEM SCIENCES, 2003, 66 (04) :671-687
[2]  
Ailon Nir, 2006, ACM S THEOR COMP STO
[3]  
Ailon Nir, 2008, ACM SIAM S DISCR ALG
[4]   Streaming Algorithms via Precision Sampling [J].
Andoni, Alexandr ;
Krauthgamer, Robert ;
Onak, Krzysztof .
2011 IEEE 52ND ANNUAL SYMPOSIUM ON FOUNDATIONS OF COMPUTER SCIENCE (FOCS 2011), 2011, :363-372
[5]  
Andoni Alexandr, 2012, HIGH FREQUENCY MOMEN
[6]  
Arora S, 2006, LECT NOTES COMPUT SC, V4110, P272
[7]  
Auerbach H., 1930, THESIS
[8]  
Avron H., 2013, ADV NEURAL INFORM PR, P2994
[9]  
Avron H., 2014, NIPS
[10]   BLENDENPIK: SUPERCHARGING LAPACK'S LEAST-SQUARES SOLVER [J].
Avron, Haim ;
Maymounkov, Petar ;
Toledo, Sivan .
SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2010, 32 (03) :1217-1236