arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2301.12965cs.LGcs.AIcs.IRcs.NAmath.NA
收藏

二次矩阵分解及其在流形学习中的应用

Quadratic Matrix Factorization with Applications to Manifold Learning

University of Toronto(多伦多大学)

Zheng Zhai, Hengchao Chen, Qiang Sun

摘要:

矩阵分解是一个流行的框架建模低秩数据矩阵。受流形学习问题的启发,本文提出了一个二次矩阵分解(QMF)框架来学习数据集所在的弯曲流形。与局部主成分分析等局部线性方法不同,QMF可以更好地利用底层流形的弯曲结构。在数学上,我们提出了一个交替最小化算法来优化QMF,并建立其理论收敛性。此外,为了避免可能的过拟合,我们提出了一种正则化的QMF算法,并讨论了如何调整其正则化参数。最后,我们阐述了如何将正则化QMF应用于流形学习问题。在合成流形学习数据集和两个真实数据集上的实验,包括MNIST手写数据集和低温电子显微镜数据集,证明了所提出的方法优于其竞争对手。

英文摘要:

Matrix factorization is a popular framework for modeling low-rank data matrices. Motivated by manifold learning problems, this paper proposes a quadratic matrix factorization (QMF) framework to learn the curved manifold on which the dataset lies. Unlike local linear methods such as the local principal component analysis, QMF can better exploit the curved structure of the underlying manifold. Algorithmically, we propose an alternating minimization algorithm to optimize QMF and establish its theoretical convergence properties. Moreover, to avoid possible over-fitting, we then propose a regularized QMF algorithm and discuss how to tune its regularization parameter. Finally, we elaborate how to apply the regularized QMF to manifold learning problems. Experiments on a synthetic manifold learning dataset and two real datasets, including the MNIST handwritten dataset and a cryogenic electron microscopy dataset, demonstrate the superiority of the proposed method over its competitors.

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。

↑