************************************************************************ GLDV--Workshop of the "Arbeitskreis Lexikographie" "GermaNet-Workshop: Applications of the German Wordnet in Theory und Practice" 09.10.--10.10.2003 Seminar fuer Sprachwissenschaft Abt. Computerlinguistik Universitaet Tuebingen Local Organization: Claudia Kunze, Lothar Lemnitzer & Andreas Wagner {kunze,lothar,wagner}@sfs.uni-tuebingen.de http://www.sfs.uni-tuebingen.de/lsd/GermaNet-Workshop/ ************************************************************************ ************************************************************************ This one-and-a-half day workshop aimed at bringing together GermaNet users (and users of other wordnets) at Tuebingen in order to present and discuss results and desiderata of their respective wordnet research and applications. We would like to thank the GLDV ("Gesellschaft fuer linguistische Datenverarbeitung" 'Association for Linguistic Data Processing') for the financial support which made the event possible without charging any conference fees. Our workshop attracted more than 40 participants from all over Germany and Switzerland, among them nine from language industry, the others from 16 different universities. The 14 papers presented with regard to different application scenarios have been complemented with two general discussion sessions. The workshop ended with setting up a commonly developed agenda in order to foster further research on GermaNet within various applications. Both organizers and participants enjoyed a very successful and stimulating workshop. The proceedings which were available on the conference will be published on-line (http://www.sfs.uni-tuebingen.de/GermaNet-Workshop/proceedings.html) as well as in a special issue of the journal 'GLDV-Forum'. ***WORKSHOP--CONTRIBUTIONS*** ---------------------------------------------------------- ***1st SESSION: Applications of GermaNet and EuroWordNet*** ---------------------------------------------------------- 1. Iman Thabet, Bernd Ludwig, Frank-Peter Schweinberger, Kerstin Buecher & Guenther Goerz (Universitaet Erlangen): "Using EWN within the Speech Operated System EMBASSI" 2. Manuela Kunze & Dietmar Roesner (Universitaet Magdeburg): "Issues in Exploiting GermaNet as a Resource in Real Applications" 3. Matthias Joerg (DaimlerChrysler AG Ulm): "Die semantische Auswertung von Produktanforderungen mit Hilfe des GermaNet" 4. Rainer Osswald (Fern-Universitaet Hagen): "Die Verwendung von GermaNet zur Pflege und Erweiterung des Computerlexikons HaGenLex" 5. Kathrin Beck (Universitaet Tuebingen): "Ein Vokabeltrainer auf der Grundlage von GermaNet und Mapa (Mapping Architecture for People's Associations)" 6. Daniela Alina Plewe (Berlin), Steffen Meschkat (Berlin) & Manfred Stede (Universitaet Potsdam): "GeneralNews - ein Metabrowser" ----------------------------------------------------------------------- ***2nd SESSION: Quantitative Approaches w.r.t. GermaNet and WordNet*** ----------------------------------------------------------------------- 7. Sabine Schulte im Walde (Brockhaus Duden Neue Medien GmbH Mannheim): "GermaNet Synsets as Selectional Preferences in Semantic Verb Clustering" 8. Andreas Wagner (Universitaet Tuebingen): "Estimating Frequency Counts in Multiple Inheritance Hierarchies" 9. Diana Steffen, Paul Buitelaar & Bogdan Sacaleanu (Consultants for Language Technology, Saarbruecken, & DFKI, Saarbruecken): "Domain Specific Sense Disambiguation with Unsupervised Methods" 10. Christian Biemann, Stefan Bordag & Uwe Quasthoff (Universitaet Leipzig): "Lernen paradigmatischer Relationen auf iterierten Kollokationen" --------------------------------------------------------------- ***3rd SESSION: Structural Aspects and Variants of Wordnets *** --------------------------------------------------------------- 11. Maren Runte, Michael Beisswenger & Angelika Storrer (Universitaet Dortmund): "Modellierung eines Terminologienetzes fuer das automatische Linking auf der Grundlage von WordNet" 12. Eva Anna Lenz, Benjamin Birkenhake & Jan Frederik Maas (Universitaeten Dortmund und Bielefeld): "Von der Erstellung bis zur Nutzung: Wortnetze als XML Topic Maps" 13. Simon Clematide (IFI-CL Universitaet Zuerich): "GermaNet und UniNet: Anknuepfen an semantische Netze" 14. Petra Steiner (ICSI Berkeley): "FrameNet und WordNet: Perspektiven fuer die Verknuepfung zweier lexikalisch--semantischer Netze" ********* ***WORKSHOP--DISCUSSION*** The workshop ended with collecting the participants' desiderata for the further development of GermaNet by the Tuebingen team as well as a common infrastructure for co-operative work. Improving the GermaNet structure depends on funding perspectives and/or intensive collaboration within the GermaNet community. * GermaNet Website will be updated regarding various aspects: - Integration of a State of the Art report on the homepage, including actual coverage of data, data formats and tools available, etc.; - GermaNet data sample for possible users should be made available on the homepage; - a pool for collecting GermaNet software (eg. extraction and visualization tools) shall be established; - Inclusion of communication facilities (newslist, mailinglist) is envisaged; - WordNet-related assignments and topics for course and diploma theses should be systematically collected. * Perpectives for co-operation w.r.t. GermaNet: - strategies for handling domain-specific or multilingual extensions of the German wordnet that are developed by other parties should be considered, particularly regarding the integration and availability of such extension modules; - Development perspectives for GermaNet towards an 'Open Source' lexicography - shared work on the standardization of wordnets should be continued on an international basis; - specific project proposals with other partners (Leipzig, Magdeburg) should be set up and submitted to various funding institutions. * Data format of GermaNet: - Dissemination of the XML-version of GermaNet data should be extended; - Development of an RDF-representation of the GermaNet data is envisaged; - GermaNet shall also be realized as relational database. * Content-related issues (for a commmon ground of discussion): - Conceptual and technical aspects of linking the general language ressource GermaNet with domain-specific extensions shall be discussed; - Improved coverage of derivations within GermaNet is planned; - The integration of function words, multi-word lexemes, and compounds has been proposed; - Further desiderata aim at the granularization of GermaNet subcategorization frames and improvement of assigning the frames to the appropriate verb senses; - an aspectual classification of verbal concepts should be developed; - the contextualization of the synonymy relation should be tested; - a critical evaluation of artificial concepts should be performed; they should be significantly reduced and, where appropriate, replaced by assigning attributes to the respective concepts and/or semantic relations; - consistency checks of taxonomic chains in GermaNet are planned; - lexical gaps should be detected and filled by a corpus-based improvement of the conceptual coverage. ********* Claudia Kunze, Lothar Lemnitzer, Andreas Wagner