Comparison of Data Selection Techniques for the Translation of Video Lectures - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Comparison of Data Selection Techniques for the Translation of Video Lectures

Résumé

For the task of online translation of scientific video lectures, using huge models is not possible. In order to get smaller and efficient models, we perform data selection. In this paper, we perform a qualitative and quantitative comparison of several data selection techniques, based on cross-entropy and infrequent n-gram criteria. In terms of BLEU, a combination of translation and language model cross-entropy achieves the most stable results. As another important criterion for measuring translation quality in our application, we identify the number of out-of-vocabulary words. Here, infrequent n-gram recovery shows superior performance. Finally, we combine the two selection techniques in order to benefit from both their strengths.
Fichier principal
Vignette du fichier
paperAMTA.pdf (251.58 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01157888 , version 1 (28-05-2015)

Identifiants

  • HAL Id : hal-01157888 , version 1

Citer

Joern Wuebker, Hermann Ney, Martínez-Villaronga Adrià, Adrià Giménez, Alfons Juan, et al.. Comparison of Data Selection Techniques for the Translation of Video Lectures. The eleventh biennial conference of the Association for Machine Translation in the Americas (AMTA-2014), AMTA, Oct 2014, Vancouver, Canada. ⟨hal-01157888⟩
109 Consultations
80 Téléchargements

Partager

Gmail Facebook X LinkedIn More