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

Download:

Current browse context:

math.AT

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

Computer Science > Machine Learning

Title: Unsupervised Features Learning for Sampled Vector Fields

Authors: Mateusz Juda
Abstract: In this paper we introduce a new approach to computing hidden features of sampled vector fields. The basic idea is to convert the vector field data to a graph structure and use tools designed for automatic, unsupervised analysis of graphs. Using a few data sets we show that the collected features of the vector fields are correlated with the dynamics known for analytic models which generates the data. In particular the method may be useful in analysis of data sets where the analytic model is poorly understood or not known.
Subjects: Machine Learning (cs.LG); Algebraic Topology (math.AT); Dynamical Systems (math.DS)
Cite as: arXiv:1911.10023 [cs.LG]
  (or arXiv:1911.10023v2 [cs.LG] for this version)

Submission history

From: Mateusz Juda [view email]
[v1] Fri, 22 Nov 2019 13:17:13 GMT (2358kb,D)
[v2] Tue, 11 Aug 2020 10:38:40 GMT (8187kb,D)

Link back to: arXiv, form interface, contact.