Towards a Theoretical Foundation for Laplacian-Based Manifold Methods

Abstract

In recent years manifold methods have attracted a considerable amount of attention in machine learning. However most algorithms in that class may be termed “manifold-motivated” as they lack any explicit theoretical guarantees. In this paper we take a step towards closing the gap between theory and practice for a class of Laplacian-based manifold methods. We show that under certain conditions the graph Laplacian of a point cloud converges to the Laplace-Beltrami operator on the underlying manifold. Theorem 1 contains the first result showing convergence of a random graph Laplacian to manifold Laplacian in the machine learning context.

Cite

Text

Belkin and Niyogi. "Towards a Theoretical Foundation for Laplacian-Based Manifold Methods." Annual Conference on Computational Learning Theory, 2005. doi:10.1007/11503415_33

Markdown

[Belkin and Niyogi. "Towards a Theoretical Foundation for Laplacian-Based Manifold Methods." Annual Conference on Computational Learning Theory, 2005.](https://mlanthology.org/colt/2005/belkin2005colt-theoretical/) doi:10.1007/11503415_33

BibTeX

@inproceedings{belkin2005colt-theoretical,
  title     = {{Towards a Theoretical Foundation for Laplacian-Based Manifold Methods}},
  author    = {Belkin, Mikhail and Niyogi, Partha},
  booktitle = {Annual Conference on Computational Learning Theory},
  year      = {2005},
  pages     = {486-500},
  doi       = {10.1007/11503415_33},
  url       = {https://mlanthology.org/colt/2005/belkin2005colt-theoretical/}
}