加权非凸非光滑低秩矩阵填充
Weighted Nonconvex Nonsmooth Low-Rank Matrix Completion
摘要: 本文利用矩阵奇异值上的
l0范数的非凸替代族来逼近秩函数，提出一种新的加权非凸非光滑最小化问题，并使用迭代加权核范数(IRNN)算法来求解该问题。实验结果表明，该方法能够很好地处理非凸非光滑问题，实现图像去噪。
Abstract:
In this paper, we propose a new weighted nonconvex nonsmooth minimization problem and use Iteratively Reweighted Nuclear Norm (IRNN) algorithm to solve the problem. It is worth mentioning that we use the nonconvex substitution family of l0-norm on the singular value of the matrix to approximate the rank function. Experimental results show that this method can deal with nonconvex nonsmooth problems well and realize image denoising.
参考文献
|
[1]
|
Toh, K. and Yun, S. (2010) An Accelerated Proximal Gradient Algorithm for Nuclear Norm Regularized Linear Least Squares Problems. Pacific Journal of Optimization, 6, 615-640.
|
|
[2]
|
Boyd, S., Parikh, N., Chu, E., Peleato, B. and Eckstein, J. (2011) Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers. Foundations and Trends® in Machine Learning, 3, 1-122. [Google Scholar] [CrossRef]
|
|
[3]
|
Mazumder, R., Hastie, T. and Tibshirani, R. (2010) Spectral Regularization Algorithms for Learning Large Incomplete Matrices. JMLR, 11, 2287-2322.
|
|
[4]
|
Gaıffas, S. and Lecue, G. (2011) Weighted Algorithms for Compressed Sensing and Matrix Completion.
arXiv:1107.1638 [cs.IT]
|
|
[5]
|
Cai, J., Candès, E. and Shen, Z. (2010) A Singular Value Thresholding Algorithm for Matrix Completion. SIAM Journal on Optimization, 20, 1956-1982. [Google Scholar] [CrossRef]
|
|
[6]
|
Beck, A. and Teboulle, M. (2009) A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems. SIAM Journal on Imaging Sciences, 2, 183-202. [Google Scholar] [CrossRef]
|