Language Processing Technologies for Slavic Languages


LATESLAV - CP92-2824

Keywords: natural language technology, grammar checking


Start Date: 1 January 93 / Status: finished / Duration: 42 months

[ participants / contact]


Objectivies and Approach

The objective of the project is to develop high level language technology for Slavic languages, in particular, grammar-checkers for Bulgarian and Czech and to provide prototypes that can be further developed by the commercial partners. Thus the research focuses on computer formalisms suitable for implementation of grammar-checkers, computationally implementable syntactic formalisms, and the syntax of Bulgarian and Czech. The implementation tasks are mainly oriented towards implementation of formalisms for syntactic analysis and error detection, development of computer grammars, and the implementation of prototypes of grammar-checkers.

Progress and Results

Within the duration of the project, work has been performed in: study of relevant differences between the investigated Slavic languages and the languages for which similar technologies already exist; design and implementation of descriptive formalisms and processing methods for free word order languages; development and implementation of an independent grammar checking formalism. The outcome was pilot implementations of grammar-checkers that were demonstrated during the project and then passed over to the industrial partners in Prague and Sofia.

Information Dissemination Activities and/or Exploitation

Besides the 16 publications during the project, the main significance of the results lies in developing and demonstrating a practically feasible technology for grammar checking of documents written in Bulgarian and Czech, in gaining new insights in the needs and ways of explicit description of the linguistic knowledge of the respective languages and in developing language-independent strategies of search for syntactic errors based on the technique of relaxation of grammatical constrains.

The work in the project has also brought along a practical spin-off that, in the short perspective, is commercially more important than the actual objective itself: an extended spelling checker for Czech, already implemented, and for Bulgarian, in implementation, that also detect trivial errors that cannot be detected by a standard spelling checker. The extension with facilities for lexical derivations is in progress.


Coordinator

Computational Linguistics
University of Saarland
Postfach 1150
D-66041 Saarbrücken, D

EU Partners

University Autonoma de Barcelona, E
University of Saarland, D

Non-EU Partners

Charles University, Prague, CZ
Marcon Ltd. Prague, CZ
Bulgarian Academy of Sciences, BG
Bulgarian Business Systems, Sofia, BG

CONTACT POINT

Prof. Dr. Hans Uszkoreit / Karel Oliva
Tel: +49 681 302 4115 / +49 681 302 4119
Fax: +49 681 302 4700
E-mail: uszkoreit@coli.uni.sbide / oliva@coli.uni.sbide


INCO synopses home page INCO cooperation index INCO keyword index
INCO acronym index INCO number index INCO Projects index
All
synopses home page all keywords index all acronyms index all numbers index

LATESLAV - CP92-2824, May 1997


please address enquiries to the ESPRIT Information Desk

html version of synopsis by Nick Cook