Terminological Enrichment for non-Interactive MT Evaluation

Abstract

In a previous study (Dabbadie, Mustafa, Timimi, 2001) we set a methodology for non interactive machine translation evaluation on big corpora, assuming that the goal of the translation was a simple understanding of the original message. The source text, in French, provided by INRA (Institut National pour la Recherche Agronomique i.e. National Institute for Agronomic Research) deals with biotechnology and animal reproduction. It has been translated into English by REVERSO. The output of the system (i.e. the result of the assembling of several components), as opposed to its individual modules or specific components (i.e. analysis, generation, grammar, lexicon, core, etc.), has been evaluated. In the present study we will recall the methodology and results obtained in the case of simple translation by REVERSO with no terminological enrichment and compare them to the results obtained after terminological  enrichment. The aim of this study is to evaluate the impact of specific terminology when integrated to an MT System and after having run the system with a basic bilingual dictionary.

Resources

Details