Centre for Computational Linguistics: Projects

PARLEX

Nationaal Fonds voor Wetenschappelijk Onderzoek
G. Adriaens

The starting point of this research is that language understanding is essentially an ``interactive'' process: different knowledge sources (morphological, syntactic, semantic and pragmatic) must be available at any time during language processing in order to arrive at a (the) correct interpretation of linguistic input. Furthermore, this research starts from the conviction that lexical elements are the entry points to the different knowledge types, rather than that this knowledge is stored in big separate modules.

These starting points call for a computational approach that puts the emphasis on a distributed knowledge representation and parallel processing. In this line of thought, different experimental systems for parallel representation and processing of language have been developed, starting from Small's Word Expert Parser model (Parallel Expert Parser 1 & 2; PEP and PEP2). As far as hardware and software are concerned, Sun workstations are used that run a programming environment for parallel Prolog (Logix).

This research has been reported on in lectures, papers and students' theses (M. Devos, M. Bourgois, W. Van Roosbroeck, J. Provoost).

• G. Adriaens, The Parallel Expert Parser: a meaning-oriented, lexically-guided, parallel-interactive model of Natural Language Understanding., Paper presented at the First International Workshop on Parsing Technologies., In: Proceedings, 1989, pp. 309-319

• G. Adriaens & U. Hahn (eds), Parallel Natural Language Processing, Ablex, Norwood, NJ, 1994, 480 pp.

CCL

Layout: webmaster@ccl.kuleuven.ac.be
Information Provider: Centrum voor Computerlinguïstiek
Comments to the Author: Geert.Adriaens@ccl.kuleuven.ac.be

      (C) Copyright 1996, CCL.
      All Rights Reserved.