Beyond CCA: Moment Matching for Multi-View Models

Abstract : We introduce three novel semi-parametric extensions of probabilistic canonical correlation analysis with identifiability guarantees. We consider moment matching techniques for estimation in these models. For that, by drawing explicit links between the new models and a discrete version of independent component analysis (DICA), we first extend the DICA cumulant tensors to the new discrete version of CCA. By further using a close connection with independent component analysis, we introduce generalized covariance matrices , which can replace the cumulant tensors in the moment matching framework, and, therefore, improve sample complexity and simplify derivations and algorithms significantly. As the tensor power method or orthogonal joint diagonalization are not applicable in the new setting, we use non-orthogonal joint diago-nalization techniques for matching the cumulants. We demonstrate performance of the proposed models and estimation techniques on experiments with both synthetic and real datasets.
Type de document :
Pré-publication, Document de travail
2016
Liste complète des métadonnées

Littérature citée [59 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01291060
Contributeur : Anastasia Podosinnikova <>
Soumis le : dimanche 20 mars 2016 - 18:48:47
Dernière modification le : jeudi 26 avril 2018 - 10:29:11
Document(s) archivé(s) le : mardi 21 juin 2016 - 10:13:43

Fichier

cca16.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01291060, version 1
  • ARXIV : 1602.09013

Collections

Citation

Anastasia Podosinnikova, Francis Bach, Simon Lacoste-Julien. Beyond CCA: Moment Matching for Multi-View Models. 2016. 〈hal-01291060〉

Partager

Métriques

Consultations de la notice

225

Téléchargements de fichiers

117