UoB : MCTS at UoB : The HIT Centre : BATMULT home


The research environments at Bergen have in recent years been involved in several large, national and international cooperative projects with relevance to multilinguality, terminology, computational linguistics and language technology. The proposed training site will offer fellows opportunities to take part in these activities. Some recent projects (2000):

SCARRIE, a project funded in part by the European Commission. The aim of the project has been to develop multilingual automatic proofreading tools for Norwegian, Swedish and Danish. This work involved cooperation with major Scandinavian publishers. 

NorGram, supported by the Norwegian Research Council but part of a larger project, ParGram, where the objective is to develop multilingual parallel grammars for Norwegian, English, French, German, Japanese, and Urdu, based on principles from Lexical-Functional Grammar. Tight cooperation with Xerox PARC, Palo Alto, and Xerox European Research Centre, Grenoble. 

From Parallel Corpus to Word Net, funded by the Meltzer Foundation, follows up the research done on Semantic Mirrors (see below). The aim is to develop a method to automatically derive semantic distinctions from translation phenomena, i.e., to use translations as empirical data for multilingual semantics. The objectively derived semantic feature structures express the kind of semantic connections between words that are represented in word nets and thesauri and are useful e.g. in connection with multilinguality, machine translation and logical inference based on natural language. 

Semantic Mirrors, a project funded in part by the Norwegian Science Council, has used parallel corpora of translated texts (consisting of source texts aligned with their translations) in order to automatically systematize the correspondence between individual words in one language and their translations into other languages. Advanced methods automatically distinguish between synonyms (different words whose translations coincide), homonyms (where the translations of a single word diverge), and other lexical relations which are crucial in detecting and preventing possible miscommunication.

PONS, research on automatic translation based on 'partial translation between related languages', is being followed up by doctoral project based on parallel texts in English and Norwegian, and aims at investigating the complexity of translation. 

TGR-SIM, a project on theoretical grammar simulation using computer methodologies. This project involves cooperation with international partners including NEC Research Institute. 

Semi-automatic Term Extraction for Norwegian, funded by the Meltzer Foundation, is a new follow-up on other terminological work going on in Bergen. In the context of terminology research, Bergen has an extensive cooperation with the Department of Computational Linguistics in Stuttgart (IMS) and the Text Laboratory and the Documentation Project at the University of Oslo, as well as with many private sector companies who have stakes in terminology development, e.g. Hydro and Statoil.


Research Areas


How to Apply


Who Where






Official page
Last updated Sept. 25, 2001 by Kristin Bech