Table of Contents
CL Web Demos
Vé - Learning by Translating
Vé is a language learning site developed by Dolores Batinić, Fabienne Leuenberger, and Kornelia Papp at the Department of Computational Linguistics of the University of Zurich. It is currently available for Spanish, but further languages will be added in the future. It is a learning-by-translating system, suitable both for beginners and advanced learners. Vé evaluates translations based on a target translation and a machine translation, and provides feedback to the user about his/her input. The system also offers on demand vocabulary help.
- Dolores Batinić: firstname.lastname@example.org
- Fabienne Leuenberger: email@example.com
- Kornelia Papp: firstname.lastname@example.org
Users of bilingwis can find translations of German or French terms within their context and compare them directly with each other. bilingwis was implemented in MySQL and PHP and is a statistical system, which makes it more flexible than a system based on a manually compiled dictionary. Even new word creations and translations which do not appear in a dictionary can be found by a statistical system.
Furthermore, the search results can be sorted by frequency which allows for conclusions about their use and spread in a certain text sort. The underlying system of bilingwis can be complemented with further text compilations in the future. Since a correct connection between the single words of the text collections cannot always be established, negative reports can occur.
Multilingwis is a web-based application for searching translations of multi-word units in large, openly available multiparallel corpora. The first edition of the tool covers the debates of the European Parliament in five languages: English, French, German, Italian, and Spanish and provides a simple and intuitive user interface, which optimally supports content-oriented queries while relieving the user from specifying complicated search expressions in a complex query language.
ParZu is a dependency parser for German. This means that it analyzes the linguistic structure of sentences and, among other things, identifies the subject and object(s) of a verb.
It is a fork of Gerold Schneider's English Pro3Gres parser. Its architecture is hybrid and consists of both a hand-written grammar and a statistics module that returns the most likely analysis of a sentence. The primary difference to the English parser is the German grammar and statistics module. Architecturally, it is different in that it supports the use of morphological information, and does not use a chunker. ParZu also has a python wrapper that supports various input/output formats and multiprocessing.