A comparison of discriminative training criteria for continuous space translation models - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Machine Translation Année : 2017

A comparison of discriminative training criteria for continuous space translation models

Résumé

This paper explores a new discriminative training procedure for continuous-space translation models (CTMs) which correlates better with translation quality than conventional training methods. The core of the method lays in the definition of a novel objective function which enables us to effectively integrate the CTM with the rest of the translation system through N-best rescoring. Using a fixed architecture, where we iteratively retrain the CTM parameters and the log-linear coefficients, we compare various ways to define and combine training criteria for each of these steps, drawing inspirations both from max-margin and learning-to-rank techniques. We experimentally show that a recently introduced loss function, which combines these two techniques, outperforms several objective functions from the literature. We also show that ensuring the consistency of the losses used to train these two sets of parameters is beneficial to the overall performance.
Fichier principal
Vignette du fichier
hal-journal-16.pdf (480.58 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01621763 , version 1 (30-10-2018)

Identifiants

  • HAL Id : hal-01621763 , version 1

Citer

Alexandre Allauzen, Quoc Khanh Do, François Yvon. A comparison of discriminative training criteria for continuous space translation models. Machine Translation, 2017, 1-2, 31, pp.19-33. ⟨hal-01621763⟩
64 Consultations
94 Téléchargements

Partager

Gmail Facebook X LinkedIn More