English

Helmholtzian Eigenmap: Topological feature discovery & edge flow learning from point cloud data

Machine Learning 2023-11-01 v3 Machine Learning

Abstract

The manifold Helmholtzian (1-Laplacian) operator Δ1\Delta_1 elegantly generalizes the Laplace-Beltrami operator to vector fields on a manifold M\mathcal M. In this work, we propose the estimation of the manifold Helmholtzian from point cloud data by a weighted 1-Laplacian L1\mathcal L_1. While higher order Laplacians have been introduced and studied, this work is the first to present a graph Helmholtzian constructed from a simplicial complex as a consistent estimator for the continuous operator in a non-parametric setting. Equipped with the geometric and topological information about M\mathcal M, the Helmholtzian is a useful tool for the analysis of flows and vector fields on M\mathcal M via the Helmholtz-Hodge theorem. In addition, the L1\mathcal L_1 allows the smoothing, prediction, and feature extraction of the flows. We demonstrate these possibilities on substantial sets of synthetic and real point cloud datasets with non-trivial topological structures; and provide theoretical results on the limit of L1\mathcal L_1 to Δ1\Delta_1.

Cite

@article{arxiv.2103.07626,
  title  = {Helmholtzian Eigenmap: Topological feature discovery & edge flow learning from point cloud data},
  author = {Yu-Chia Chen and Weicheng Wu and Marina Meilă and Ioannis G. Kevrekidis},
  journal= {arXiv preprint arXiv:2103.07626},
  year   = {2023}
}
R2 v1 2026-06-24T00:05:53.466Z