Intelligent Database Systems Lab N.Y.U.S. T. I. M. Word sense disambiguation of WordNet glosses Presenter: Chun-Ping Wu Author: Dan Moldovan, Adrian Novischi Computer Speech and Language, 2004 國國國國國國國國 National Yunlin University of Science and Technology 2011/02/10
Word sense disambiguation of WordNet glosses. Presenter: Chun-Ping Wu Author: Dan Moldovan, Adrian Novischi. 國立雲林科技大學 National Yunlin University of Science and Technology. 2011/02/10. Computer Speech and Language, 2004. Outline. Motivation Objective WordNet Methodology Experiments - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.
Word sense disambiguation of WordNet glosses
Presenter: Chun-Ping Wu Author: Dan Moldovan, Adrian Novischi
Computer Speech and Language, 2004
國立雲林科技大學National Yunlin University of Science and Technology
2011/02/10
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Outline Motivation Objective WordNet Methodology Experiments Conclusion Comments
2
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Motivation
Manual disambiguation is known to be very laborious and time intensive.
It’s difficult to obtain a semantically tagged corpus and the features appearing in such corpus are very sparse, machine learning techniques were not found to be very successful.
3
This is my watch.(手錶 ?注視 ?)
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Objective To present a suite of methods and results for the semantic
disambiguation of WordNet glosses.
4
This is my watch.(手錶 )
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.WordNet
WordNet Noun
ISA relation Verb
Change, communication, cognition, creation, emotion, etc.
Adjective Synonym/Antonym
Adverb Synonym/Antonym
5
gloss
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Methodology Semantic disambiguation methods
Monosemous words Same hierarchy relation Lexical parallelism SemCor bigrams Cross-reference Reversed cross-reference Distance among glosses Common domain Patterns First sense restricted Building the WSD system using the methods
6
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Methodology Monosemous words
Same hierarchy relation
7
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Methodology Lexical parallelism
SemCor bigrams
8
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Methodology Cross-reference
Reversed cross-reference
9
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Methodology Distance among glosses
Common domain
10
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Methodology Patterns
11
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Methodology First sense restricted
A sense of noun or verb is more general if it has the smallest number of ancestors from all senses in the ISA hierarchy.
A sense of an adjective is more general if it has the largest number of similarity pointers from all senses.
Building the WSD system using the methods XWN_WSD
12
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Experiments
Contribution of each method
13
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Experiments
14
Voting
Intelligent Database Systems Lab
N.Y.U.S.T.
I. M.Conclusion
1515
A suite of heuristical methods are presented for the disambiguation of WordNet glosses.
Once the WordNet glosses are disambiguated, several applications become possible. QA System