Document Translation for Cross-Language Text Retrieval at the University of Maryland

Evaluation d'une technique de recherche de texte, par croisement de langue, basee sur une traduction entierement automatique, based on les approches basees sur the traduction du document.

Abstract

The University of Maryland participated in three TREC-6 tasks: ad hoc retrieval, cross-language retrieval, and spoken document retrieval. The principal focus of the work was evaluation of a cross-language text retrieval technique based on fully automatic machine translation. The results show that approaches based on document translation can be approximately as effective as approaches based on query translation, but that additional work will be needed to develop a solid basis for choosing between the two in specific applications. Ad hoc and spoken document retrieval results are also presented.