作者: Philippe Langlais , Stéphane Huet , Julien Bourdaillet
DOI:
关键词: Computer science 、 Service (systems architecture) 、 Artificial intelligence 、 Natural language processing 、 Concordancer 、 Engineering drawing 、 Computer-assisted translation 、 Machine translation
摘要: Despite the impressive amount of recent studies devoted to improving state art machine translation, computer assisted translation tools remain preferred solu- tion human translators when publication quality is concern. In this paper, we present our ongoing efforts conducted within a project which aims at commercial bilingual concordancer TransSearch. The core technology Web-based service mainly relies on sentence-level alignment. study, discuss and evaluate embedding statistical word-level Two novel issues that are essential success new prototype tackled: detecting erroneous translations grouping together similar translations. Mots-cles : alignement au niveau des mots, concordancier bilingue, traduction automa- tique.