We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Shape-Preserving Dimensionality Reduction : An Algorithm and Measures of Topological Equivalence

Abstract: We introduce a linear dimensionality reduction technique preserving topological features via persistent homology. The method is designed to find linear projection $L$ which preserves the persistent diagram of a point cloud $\mathbb{X}$ via simulated annealing. The projection $L$ induces a set of canonical simplicial maps from the Rips (or \v{C}ech) filtration of $\mathbb{X}$ to that of $L\mathbb{X}$. In addition to the distance between persistent diagrams, the projection induces a map between filtrations, called filtration homomorphism. Using the filtration homomorphism, one can measure the difference between shapes of two filtrations directly comparing simplicial complexes with respect to quasi-isomorphism $\mu_{\operatorname{quasi-iso}}$ or strong homotopy equivalence $\mu_{\operatorname{equiv}}$. These $\mu_{\operatorname{quasi-iso}}$ and $\mu_{\operatorname{equiv}}$ measures how much portion of corresponding simplicial complexes is quasi-isomorphic or homotopy equivalence respectively. We validate the effectiveness of our framework with simple examples.
Comments: 18 pages, 2 figures
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:2106.02096 [stat.ML]
  (or arXiv:2106.02096v3 [stat.ML] for this version)

Submission history

From: Kisung You [view email]
[v1] Thu, 3 Jun 2021 19:28:52 GMT (1775kb)
[v2] Mon, 7 Jun 2021 16:59:07 GMT (1775kb)
[v3] Sun, 13 Jun 2021 21:03:10 GMT (1775kb)

Link back to: arXiv, form interface, contact.