European Commission logo
français français
CORDIS - Résultats de la recherche de l’UE
CORDIS
Contenu archivé le 2024-06-18

Testing the portability of techniques to handle dissimilar source and target languages in MT

Objectif

"The aim of the ENEUS project is to develop an English-Basque machine translation system which responds to two main objectives: (1) it contributes to the MT research field by testing the portability of techniques to handle dissimilar source and target languages; and (2) it meets the social demand for a good quality English-Basque translation system.
Both rule-based and statistical approaches will be explored based on the systems and results obtained in the OpenMT (2006-2009) and EUSMT (Labaka, 2009) projects developed by the Ixa Group to translate from Spanish to Basque.
Matxin, an open rule-based machine translation system, will be extended to the English/Basque language pair. A hybrid constituent and dependency-based scheme for the analysis of Spanish will be transferred to English, integrating the successful Freeling Suite of Language Analyzers; the language pair structural differences will be thoroughly explored to fine-tune the existing limited transfer rules; and the maximation of reuse regarding the existing language pair will be exploited.
Based on EUSMT, a customised Moses baseline to improve segmentation and reordering for dissimilar languages and Basque in particular, the ENEUS project will investigate the statistical independence for the morphemes measured as χ2 on a large monoligual corpus to adapt the optimal segmentation option to the English-Basque language pair. Simultaneously the set of reordering rules will be adapted to the new language pair.
During the project, an effort to extend the bilingual and monolingual corpora is also planned given its importance to exploit the latest corpus-based approaches.
The ENEUS project brings together expertise from three different fields, that of linguistics, computer science and translation, allowing the fellow to re-start her research career within a reknown NLP research group receiving specialised computational training while contributing with a specialised linguistic and translation background."

Appel à propositions

FP7-PEOPLE-2011-IEF
Voir d’autres projets de cet appel

Coordinateur

UNIVERSIDAD DEL PAIS VASCO/ EUSKAL HERRIKO UNIBERTSITATEA
Contribution de l’UE
€ 168 896,40
Adresse
BARRIO SARRIENA S N
48940 Leioa
Espagne

Voir sur la carte

Région
Noreste País Vasco Bizkaia
Type d’activité
Higher or Secondary Education Establishments
Contact administratif
Charo Sánchez (Ms.)
Liens
Coût total
Aucune donnée