Space-time is a profound concept in physics. This concept was shown to be useful for dimensionality reduction. We present basic definitions with interesting counter-intuitions. We give theoretical propositions to show that space-time is a more powerful representation than Euclidean space. We apply this concept to manifold learning for preserving local information. Empirical results on non-metric datasets show that more information can be preserved in space-time.
K Sun (University of Geneva)
Jun Wang (Expedia, Geneva)
Stephane Marchand-Maillet (University of Geneva)
More from the Same Authors
2018 Poster: Representation Learning of Compositional Data »
Marta Avalos · Richard Nock · Cheng Soon Ong · Julien Rouar · Ke Sun