Cartan Networks: Group theoretical Hyperbolic Deep Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Milanesio, Federico, Santoro, Matteo, Fré, Pietro G., Sanguinetti, Guido
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910975264817152
author Milanesio, Federico
Santoro, Matteo
Fré, Pietro G.
Sanguinetti, Guido
author_facet Milanesio, Federico
Santoro, Matteo
Fré, Pietro G.
Sanguinetti, Guido
contents Hyperbolic deep learning leverages the metric properties of hyperbolic spaces to develop efficient and informative embeddings of hierarchical data. Here, we focus on the solvable group structure of hyperbolic spaces, which follows naturally from their construction as symmetric spaces. This dual nature of Lie group and Riemannian manifold allows us to propose a new class of hyperbolic deep learning algorithms where group homomorphisms are interleaved with metric-preserving diffeomorphisms. The resulting algorithms, which we call Cartan networks, show promising results on various benchmark data sets and open the way to a novel class of hyperbolic deep learning architectures.
format Preprint
id arxiv_https___arxiv_org_abs_2505_24353
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Cartan Networks: Group theoretical Hyperbolic Deep Learning
Milanesio, Federico
Santoro, Matteo
Fré, Pietro G.
Sanguinetti, Guido
Machine Learning
Hyperbolic deep learning leverages the metric properties of hyperbolic spaces to develop efficient and informative embeddings of hierarchical data. Here, we focus on the solvable group structure of hyperbolic spaces, which follows naturally from their construction as symmetric spaces. This dual nature of Lie group and Riemannian manifold allows us to propose a new class of hyperbolic deep learning algorithms where group homomorphisms are interleaved with metric-preserving diffeomorphisms. The resulting algorithms, which we call Cartan networks, show promising results on various benchmark data sets and open the way to a novel class of hyperbolic deep learning architectures.
title Cartan Networks: Group theoretical Hyperbolic Deep Learning
topic Machine Learning
url https://arxiv.org/abs/2505.24353