[Conference Home Page]

[Conference Home Page]

Lexicons are indispensable resources for almost every natural language project. To date, WordNet 1.6 represents the largest publicly available on-line lexical resource, already used in various applications of the human language technology. Systems performing word sense disambiguation, information extraction or retrieval, prepositional attachment, interpretation of nominalizations, textual summarization, coreference resolution, abductive reasoning conversational implicature, recognition of textual cohesion and coherence, intelligent Internet searches and some of the digital libraries projects use WordNet.

This workshop intends to bring together researchers that use WordNet in different systems and to focus on two particular issues: (a) how to customize the knowledge derived from WordNet for various NLP applications and (b) how to derive methods that infer semantic information using WordNet. The contributions might address one or more of the following questions:
 
 

• What are the NLP applications for which WordNet is a valuable resource and how much effort was involved to integrate it in your systems?
• Is WordNet used to build ad-hoc ontologies? What are the applications that use WordNet-derived ontologies?
• How can WordNet be used to develop a word sense disambiguation algorithm of high performance?
• How to extend WordNet for identifying thematic roles and resolving verb polysemy?
• What minimal customization should be implemented to use WordNet for a large-scale abductive reasoning system?
• Is WordNet a lexical knowledge base that can be easily used to adjust Information Extraction systems across domains?
• Are the lexico-semantic relations from WordNet a valid base for developing an extended coreference task for information extraction, and what are the possible methodologies?
• How can WordNet be mined to find textual implied information and what is the degree of plausibility of the returned information?
• What are the approaches of using the extensive linguistic knowledge of WordNet to derive the discourse structure of a text; can it be the only knowledge source and if not, what additional knowledge may be used?
• What is the current performance boost provided by WordNet in the systems using it? Could your systems perform without WordNet?
• What are the desirable features of WordNet for your system, and what would be the predicted performance increase when having them?

Organizing committee

The workshop is organized by

Sanda Harabagiu (SRI International)
Joyce Yue Chai (Duke University)

Program committee

• Alan Biermann (Duke University)
• Joyce Chai (Duke University)
• Martin Chodorow (New York University)
• Christiane Fellbaum (Princeton University)
• Fernando Gomez (University of Central Florida)
• Ken Haase (MIT)
• Sanda Harabagiu (SRI International)
• Marti Hearst (University of California, Berkeley)
• Graeme Hirst (University of Toronto)
• Claudia Leacock (Educational Testing Service)
• Mitch Marcus (University of Pennsylvania)
• George A. Miller (Princeton University)
• Dan Moldovan (Southern Methodist University)
• Hwee Tou Ng (DSO National Laboratories, Singapore)
• Philip Resnik (University of Maryland)
• Yorick Wilks (University of Sheffield)

Sanda Harabagiu
SRI International 
333 Ravenswood Ave 
Menlo Park, CA 94025 
U.S.A. 
(Ph) (650) 859-3852 
(Fax) (650) 859-3735 
harabagi@ai.sri.com