Cluster indicator decomposition for efficient matrix factorization

Authors:
Dijun Luo;Chris Ding;Heng Huang
Affiliations:
Computer Science and Engineering Department, University of Texas at Arlington;Computer Science and Engineering Department, University of Texas at Arlington;Computer Science and Engineering Department, University of Texas at Arlington
Venue:
IJCAI'11 Proceedings of the Twenty-Second international joint conference on Artificial Intelligence - Volume Volume Two
Year:
2011

Citing 9
Cited 0

SIMPLIcity: Semantics-Sensitive Integrated Matching for Picture LIbraries

IEEE Transactions on Pattern Analysis and Machine Intelligence
Pass efficient algorithms for approximating large matrices

SODA '03 Proceedings of the fourteenth annual ACM-SIAM symposium on Discrete algorithms
Spectral Grouping Using the Nyström Method

IEEE Transactions on Pattern Analysis and Machine Intelligence
Learning Generative Visual Models from Few Training Examples: An Incremental Bayesian Approach Tested on 101 Object Categories

CVPRW '04 Proceedings of the 2004 Conference on Computer Vision and Pattern Recognition Workshop (CVPRW'04) Volume 12 - Volume 12
Fast monte-carlo algorithms for finding low-rank approximations

Journal of the ACM (JACM)
Predictive low-rank decomposition for kernel methods

ICML '05 Proceedings of the 22nd international conference on Machine learning
Orthogonal nonnegative matrix t-factorizations for clustering

Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data mining
Improved Nyström low-rank approximation and error analysis

Proceedings of the 25th international conference on Machine learning
Non-negative Laplacian Embedding

ICDM '09 Proceedings of the 2009 Ninth IEEE International Conference on Data Mining

Quantified Score

Hi-index	0.00

Visualization

Abstract

We propose a new clustering based low-rank matrix approximation method, Cluster Indicator Decomposition (CID), which yields more accurate low-rank approximations than previous commonly used singular value decomposition and other Nyström style decompositions. Our model utilizes the intrinsic structures of data and theoretically be more compact and accurate than the traditional low rank approximation approaches. The reconstruction in CID is extremely fast leading to a desirable advantage of our method in large-scale kernel machines (like Support Vector Machines) in which the reconstruction of the kernels needs to be frequently computed. Experimental results indicate that our approach compress images much more efficiently than other factorization based methods. We show that combining our method with Support Vector Machines obtains more accurate approximation and more accurate prediction while consuming much less computation resources.