Spoken Language Translation Graphs Re-decoding using Automatic Quality Assessment

This paper investigates how automatic quality assessment of spoken language translation (SLT) can help re-decoding SLT output graphs and improving the overall speech translation performance. Using robust word confidence measures (from both ASR and MT) to re-decode the SLT graph leads to a significant BLEU improvement (more than 2 points) compared to our SLT baseline (French-English task).

Domaines

Informatique et langage [cs.CL]

Fichier principal

poster_ASRU2015.pdf (1.01 Mo)

Origine	Fichiers produits par l'(les) auteur(s)

Benjamin Lecouteux : Connectez-vous pour contacter le contributeur

https://hal.science/hal-02095256

Soumis le : mercredi 10 avril 2019-12:11:20

Dernière modification le : lundi 16 décembre 2024-15:38:04

Dates et versions

hal-02095256 , version 1 (10-04-2019)

Identifiants

HAL Id : hal-02095256 , version 1

Citer

Laurent Besacier, Benjamin Lecouteux, Ngoc Quang Luong, Ngoc Tien Le. Spoken Language Translation Graphs Re-decoding using Automatic Quality Assessment. ASRU, 2015, Scotsdale, United States. ⟨hal-02095256⟩

Exporter

BibTeX XML-TEI Dublin Core DC Terms EndNote DataCite

Collections

UGA CNRS LIG LIG_TDCGE_GETALP POLYTECH-GRENOBLE LIG_SIDCH

44 Consultations

92 Téléchargements