Nowadays many high-dimensional data in real-world applications can be modeled as data points lying close to a low-dimensional nonlinear manifold. However, discovering the structure of the manifold from a set of data points sampled from the manifold possibly with noise is a very challenging unsupervised learning[3]. Traditional dimension reduction techniques work well when the data points lie close to a linear subspace, however, it failed to detect nonlinear structures of the set of data points, such as principal component analysis and factor analysis. To solve this problem, there have been much renewed interests in developing efficient algorithms for constructing nonlinear low-dimensional manifolds from sample data points in high-dimensional manifold, emphasizing simple algorithmic implementation and avoiding optimization problems prone to local minimal[4]-[7], nevertheless many fundamental problems are required to be further investigated.