How can shallow NLP help a machine translation system

Petr Homola, Jakub Piskorski

In: Proceedings of the Conference Human Language Technologies - The Baltic Perspective, April 2004. Annual Meeting of the Association for Computational Linguistics: Human Language Technologies (ACL-HLT) Riga, Latvia 2004.


The historical EU enlargement will have an enormous impact on all European countries. In particular, due to the wide variety of languages spoken in the extended EU, machine translation (MT) poses a challenging and intriguing task. Six from the �new� EU languages belong to the Balto-Slavonic language family. This paper focuses on MT among these languages, which can be achieved by relatively simple means. We present an experimental MT system for related languages (currently Baltic and Slavonic) and explain how its complexity can be reduced by exploiting similarities between the source and target language.

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence