Homología persistente en 3-variedades
Loading...
Identifiers
Publication date
Authors
Advisors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Se comienza con una introducción al análisis topológico de datos, presentando las herramientas básicas de esta disciplina: homologíaa persistente, diagramas de persistencia y landscapes, teorema de estabilidad, entre otras. En particular, nos interesará la aplicación de la homología persistente sobre muestras finitas de puntos sobre una variedad, a partir de las construcciones de complejos simpliciales sobre dichas muestras, debido a su potencial para generar invariantes métricos. Se sigue con los fundamentos de la geometría hiperbólica, estudiando algunos de los modelos clásicos del espacio hiperbólico, el cual es cubierta universal de cualquier variedad hiperbólica, particularizando el caso de dimensión 3. El objetivo principal de esta sección es demostrar el Teorema de Rigidez de Mostow en el caso compacto; para ello, se muestran previamente algunos conceptos y resultados usados en la demostración. Como consecuencia de este teorema, en el caso de 3-variedades hiperbólicas, la métrica es un invariante topológico y por lo tanto, variedades hiperbólicas no homeomorfas pueden ser distinguidas usando invariantes métricos como los que nos proporcionará la homología persistente. Finalmente, se llevarán las técnicas anteriores a la práctica, con la realización de un programa que muestree puntos aleatorios sobre 3-variedades hiperbólicas compactas orientables, calcule los correspondientes diagramas de persistencia y landscapes, y compare mediante contrastes de hipótesis los resultados obtenidos para cualquier par de variedades hiperbólicas dadas, con el objetivo de distinguirlas topológicamente con cierto grado de confianza.
We begin with an introduction to topological data analysis, presenting the basic tools of this discipline: persistent homology, persistence diagrams and landscapes, as well as the stability theorem, among others. Specifically, we are interested in the application of persistent homology to finite samples of points on a manifold, using the construction of simplicial complexes on these samples, due to their potential to generate metric invariants. Next, we cover the fundamentals of hyperbolic geometry, studying some of the classical models of hyperbolic space, which is the universal cover of any hyperbolic manifold, focusing on the 3- dimensional case. The main objective of this section is to demonstrate Mostow’s Rigidity Theorem in the compact case. To this end, we present some concepts and results used in the proof. As a consequence of this theorem, in the case of 3-dimensional hyperbolic manifolds, the metric is a topological invariant. Therefore, non-homeomorphic hyperbolic manifolds can be distinguished using metric invariants, such as those provided by persistent homology. Finally, we will put the previous techniques into practice by developing a program that samples random points on compact orientable 3-dimensional hyperbolic manifolds, calculates the corresponding persistence diagrams and landscapes, and compares the results obtained for any pair of given hyperbolic manifolds using hypothesis testing, with the aim of topologically distinguishing them with a certain degree of confidence.
We begin with an introduction to topological data analysis, presenting the basic tools of this discipline: persistent homology, persistence diagrams and landscapes, as well as the stability theorem, among others. Specifically, we are interested in the application of persistent homology to finite samples of points on a manifold, using the construction of simplicial complexes on these samples, due to their potential to generate metric invariants. Next, we cover the fundamentals of hyperbolic geometry, studying some of the classical models of hyperbolic space, which is the universal cover of any hyperbolic manifold, focusing on the 3- dimensional case. The main objective of this section is to demonstrate Mostow’s Rigidity Theorem in the compact case. To this end, we present some concepts and results used in the proof. As a consequence of this theorem, in the case of 3-dimensional hyperbolic manifolds, the metric is a topological invariant. Therefore, non-homeomorphic hyperbolic manifolds can be distinguished using metric invariants, such as those provided by persistent homology. Finally, we will put the previous techniques into practice by developing a program that samples random points on compact orientable 3-dimensional hyperbolic manifolds, calculates the corresponding persistence diagrams and landscapes, and compares the results obtained for any pair of given hyperbolic manifolds using hypothesis testing, with the aim of topologically distinguishing them with a certain degree of confidence.
Description
89 páxinas
Keywords
Bibliographic citation
Relation
Has part
Has version
Is based on
Is part of
Is referenced by
Is version of
Requires
Sponsors
Rights
Attribution-NonCommercial-ShareAlike 4.0 International








