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

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

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

Statistics > Machine Learning

Title: Manifold structure in graph embeddings

Abstract: Statistical analysis of a graph often starts with embedding, the process of representing its nodes as points in space. How to choose the embedding dimension is a nuanced decision in practice, but in theory a notion of true dimension is often available. In spectral embedding, this dimension may be very high. However, this paper shows that existing random graph models, including graphon and other latent position models, predict the data should live near a much lower-dimensional set. One may therefore circumvent the curse of dimensionality by employing methods which exploit hidden manifold structure.
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:2006.05168 [stat.ML]
  (or arXiv:2006.05168v3 [stat.ML] for this version)

Submission history

From: Patrick Rubin-Delanchy Dr [view email]
[v1] Tue, 9 Jun 2020 10:30:17 GMT (3688kb,D)
[v2] Wed, 28 Oct 2020 16:21:11 GMT (2217kb,D)
[v3] Tue, 5 Jan 2021 11:17:07 GMT (2217kb,D)

Link back to: arXiv, form interface, contact.