Distance Measure Machines

Abstract : This paper presents a distance-based discriminative framework for learning with probability distributions. Instead of using kernel mean embeddings or generalized radial basis kernels, we introduce embeddings based on dissimilarity of distributions to some reference distributions denoted as templates. Our framework extends the theory of similarity of \citet{balcan2008theory} to the population distribution case and we show that, for some learning problems, some dissimilarity on distribution achieves low-error linear decision functions with high probability. Our key result is to prove that the theory also holds for empirical distributions. Algorithmically, the proposed approach consists in computing a mapping based on pairwise dissimilarity where learning a linear decision function is amenable. Our experimental results show that the Wasserstein distance embedding performs better than kernel mean embeddings and computing Wasserstein distance is far more tractable than estimating pairwise Kullback-Leibler divergence of empirical distributions.
Type de document :
Pré-publication, Document de travail
2018
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01717940
Contributeur : Alain Rakotomamonjy <>
Soumis le : mardi 27 février 2018 - 22:09:10
Dernière modification le : jeudi 15 novembre 2018 - 11:59:00
Document(s) archivé(s) le : lundi 28 mai 2018 - 18:30:45

Fichiers

WDMM.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01717940, version 1
  • ARXIV : 1803.00250

Citation

Alain Rakotomamonjy, Abraham Traore, Maxime Berar, Rémi Flamary, Nicolas Courty. Distance Measure Machines. 2018. 〈hal-01717940v1〉

Partager

Métriques

Consultations de la notice

535

Téléchargements de fichiers

142