arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1709.04836stat.MLcs.CVcs.LG
收藏

带特征的信息非凸鲁棒主成分分析

Informed Non-convex Robust Principal Component Analysis with Features

Imperial College London(帝国理工学院)

Niannan Xue, Jiankang Deng, Yannis Panagakis, Stefanos Zafeiriou

摘要:

我们重新审视的问题,强大的主成分分析的功能作为先验信息。为此,提出了一种新颖的,优雅的,非凸的优化方法,将给定的观测矩阵分解成一个低秩的核心和相应的稀疏残差。严格的理论分析所提出的算法的结果在准确的恢复保证低计算复杂度。适当设计的合成实验表明,我们的方法是第一个完全利用凸性的非凸性在可恢复性和速度方面的权力。也就是说,所提出的非凸方法是更准确,更快的最佳算法相比,研究中的问题。两个实际应用,即图像分类和人脸去噪进一步证明了所提出的方法的实用优势。

英文摘要:

We revisit the problem of robust principal component analysis with features acting as prior side information. To this aim, a novel, elegant, non-convex optimization approach is proposed to decompose a given observation matrix into a low-rank core and the corresponding sparse residual. Rigorous theoretical analysis of the proposed algorithm results in exact recovery guarantees with low computational complexity. Aptly designed synthetic experiments demonstrate that our method is the first to wholly harness the power of non-convexity over convexity in terms of both recoverability and speed. That is, the proposed non-convex approach is more accurate and faster compared to the best available algorithms for the problem under study. Two real-world applications, namely image classification and face denoising further exemplify the practical superiority of the proposed method.

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。

↑