SemEval 2014 Task 5 - L2 Writing Assistant

Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014), pages 36–44, Dublin, Ireland, August 23-24, 2014. SemEval-2014 Task 5: L2 Writing Assistant Maarten van Gompel, Iris Hendrickx, Antal van den Bosch Centre for Language Studies, Radboud University Nijmegen, The Netherlands proycon@anaproy.nl, i.hendrickx@let.ru.nl, a.vandenbosch@let.ru.nl Els Lefever and V´eronique Hoste LT3, Language and Translation Technology Team, Ghent University, Belgium els.lefever@ugent.be, veronique.hoste@ugent.be Abstract We present a new cross-lingual task for SemEval concerning the translation of L1 fragments in an L2 context. The task is at the boundary of Cross-Lingual Word Sense Disambiguation and Machine Translation. It finds its application in the field of computer-assisted translation, par- ticularly in the context of second language learning. Translating L1 fragments in an L2 context allows language learners when writing in a target language (L2) to fall back to their native language (L1) when- ever they are uncertain of the right word or phrase. 1 Introduction We present a new cross-lingual and application- oriented task for SemEval that is situated in the area where Word Sense Disambiguation and Ma- chine Translation meet. Finding the proper trans- lation of a word or phrase in a given context is much like the problem of disambiguating between multiple senses. In this task participants are asked to build a translation/writing assistance system that trans- lates specifically marked L1 fragments in an L2 context to their proper L2 translation. This type of translation can be applied in writing assistance systems for language learners in which users write in a target language, but are allowed to occasion- ally back off to their native L1 when they are un- certain of the proper lexical or grammatical form in L2. The task concerns the NLP back-end rather than any user interface. Full-on machine translation typically concerns the translation of complete sentences or texts from This work is licensed under a Creative Commons Attribution 4.0 International Licence: http://creativecommons.org/licenses/by/4.0/ L1 to L2. This task, in contrast, focuses on smaller fragments, side-tracking the problem of full word reordering. We focus on the following language combi- nations of L1 and L2 pairs: English-German, English-Spanish, French-English and Dutch- English. Task participants could participate for all language pairs or any subset thereof. 2 Task Description We frame the task in the context of second lan- guage learning, yielding a specific practical appli- cation. Participants build a translation assistance sys- tem rather than a full machine translation system. The L1 expression, a word or phrase, is translated by the system to L2, given the L2 context already present, including right-side context if available. The aim here, as in all translation, is to carry the semantics of the L1 fragment over to L2 and find the most suitable L2 expression given the already present L2 context. Other than a limit on length (6 words), we do not pose explicit constraints on the kinds of L1 fragments allowed. The number of L1 fragments is limited to one fragment per sentence. The task addresses both a core problem of WSD, with cross-lingual context, …

Content truncated to 500 words