Word Representations in Factored Neural Machine Translation

Abstract : Translation into a morphologically rich language requires a large output vocabulary to model various morphological phenomena, which is a challenge for neural machine translation architectures. To address this issue, the present paper investigates the impact of having two output factors with a system able to generate separately two distinct representations of the target words. Within this framework, we investigate several word representations that correspond to different distributions of morpho-syntactic information across both factors. We report experiments for translation from English into two morphologically rich languages, Czech and Latvian, and show the importance of explicitly modeling target morphology.
Type de document :
Communication dans un congrès
Association for Computational Linguistics. Conference on Machine Translation, Sep 2017, Copenhagen, Denmark. Proceedings of the Conference on Machine Translation (WMT),, 1, pp.43 - 55
Liste complète des métadonnées

Littérature citée [30 références]  Voir  Masquer  Télécharger

https://hal.archives-ouvertes.fr/hal-01618384
Contributeur : Dev.Limsi Dev.Limsi <>
Soumis le : mardi 13 février 2018 - 10:11:15
Dernière modification le : mardi 20 novembre 2018 - 14:04:02
Document(s) archivé(s) le : lundi 14 mai 2018 - 12:33:32

Fichier

WMT03.pdf
Fichiers éditeurs autorisés sur une archive ouverte

Identifiants

  • HAL Id : hal-01618384, version 1

Citation

Franck Burlot, Mercedes Garcia-Martinez, Loïc Barrault, Fethi Bougares, François Yvon. Word Representations in Factored Neural Machine Translation. Association for Computational Linguistics. Conference on Machine Translation, Sep 2017, Copenhagen, Denmark. Proceedings of the Conference on Machine Translation (WMT),, 1, pp.43 - 55. 〈hal-01618384〉

Partager

Métriques

Consultations de la notice

221

Téléchargements de fichiers

29