Bootstrapping a Statistical Speech Translator From a Rule-Based One

Authors

  • Manny Rayner Geneva University
  • Pierrette Bouillon Geneva University
  • Paula Estrella FaMAF, U. Nacional de C´ordoba
  • Yukie Nakao Département Communication, Langues & Entreprise, École Centrale de Nantes
  • Gwen Christian

DOI:

https://doi.org/10.33011/lilt.v5i.1231

Keywords:

translation, speech technology, spoken language

Abstract

We describe a series of experiments in which we start with English to French and English to Japanese versions of a rule-based speech translation system for a medical domain, and bootstrap corresponding statistical systems. Comparative evaluation reveals that the statistical systems are still slightly inferior to the rule-based ones, despite the fact that considerable effort has been invested in tuning both the recognition and translation components; however, a hybrid system is able to deliver a small but significant improvement in performance. In conclusion, we suggest that the hybrid architecture we describe potentially allows construction of limited-domain speech translation systems which combine substantial source-language coverage with high-precision translation.

Downloads

Published

2012-02-01

How to Cite

Rayner, M., Bouillon, P., Estrella, P., Nakao, Y., & Christian, G. (2012). Bootstrapping a Statistical Speech Translator From a Rule-Based One. Linguistic Issues in Language Technology, 5. https://doi.org/10.33011/lilt.v5i.1231

Issue

Section

Articles