UMIACS Computational Linguistics Colloquium Series,
October 9, 1997
UMIACS Computational Linguistics Colloquium Series,
October 9, 1997
`Raise Duty'' vs. ``Perform Duty''---Using Syntactic
Dependencies to Resolve Semantic Ambiguities
Dekang Lin
University of Manitoba
Most previous corpus-based algorithms disambiguate a word with a
classifier trained from previous usages of the same word. Separate
classifiers have to be trained for different words. I present an
algorithm that uses the same knowledge sources to disambiguate
different words. The algorithm does not require a sense-tagged
corpus and exploits the fact that two different words are likely to
have similar meanings if they occur in identical local contexts.
I will also present an information-theoretic measure of similarity
that overcomes several problems with previous similarity measures.
Links to Dr. Lin's research
interests and papers
Return to the UMD
Computational Linguistics Colloquium Series.