Hyperbolic Sliced-Wasserstein via Geodesic and Horospherical Projections - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

Hyperbolic Sliced-Wasserstein via Geodesic and Horospherical Projections

Résumé

It has been shown beneficial for many types of data which present an underlying hierarchical structure to be embedded in hyperbolic spaces. Consequently, many tools of machine learning were extended to such spaces, but only few discrepancies to compare probability distributions defined over those spaces exist. Among the possible candidates, optimal transport distances are well defined on such Riemannian manifolds and enjoy strong theoretical properties, but suffer from high computational cost. On Euclidean spaces, sliced-Wasserstein distances, which leverage a closed-form of the Wasserstein distance in one dimension, are more computationally efficient, but are not readily available on hyperbolic spaces. In this work, we propose to derive novel hyperbolic sliced-Wasserstein discrepancies. These constructions use projections on the underlying geodesics either along horospheres or geodesics. We study and compare them on different tasks where hyperbolic representations are relevant, such as sampling or image classification.

Dates et versions

hal-04356781 , version 1 (20-12-2023)

Identifiants

Citer

Clément Bonet, Laetitia Chapel, Lucas Drumetz, Nicolas Courty. Hyperbolic Sliced-Wasserstein via Geodesic and Horospherical Projections. TAG-ML 2023 ICML Workshop, Jul 2023, Honololu, Hawaii, United States. pp.3995. ⟨hal-04356781⟩
31 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More