WordNet

1
WordNet

• WordNet, WSD

2
WordNet

• What is WordNet?
• Miller 95 WordNet is an online lexical
  database designed for use under program control.
  English nouns, verbs, adjectives, and adverbs are
  organized into sets of synonyms, each
  representing a lexicalized concept. Semantic
  relations link the synonym sets.

3
WordNet

• Go to the main WordNet site
• http//wordnet.princeton.edu/
• Open the wordnet folder on pongo
• /dropbox/570/wordnet/dict

4
WordNet Vocabulary

• See glossary at http//wordnet.princeton.edu/glo
  ss
• synset A synonym set a set of words that are
  interchangeable in some context
• lemma lower case ASCII text of word as found in
  the WordNet database index files
• lexical pointer A lexical pointer indicates a
  relation between words in synsets

5
Navigating WordNet files

• data. files the actual network files (synsets)
• index. files contains lower case instances of
  all words in WordNet, with pointers to the synset
  entries in the network

6
WordNet data file
Synset file offset
Synset type
File number
words in synset
word

• 00045430 04 n 01 performance 3 003 _at_ 00033580 n
  0000 00045680 n 0000 00045874 n 0000 any
  recognized accomplishment "they admired his
  performance under stress
• 00045680 04 n 01 overachievement 0 003 _at_ 00045430
  n 0000 02537922 v 0101 ! 00045874 n 0101
  better than expected performance (better than
  might have been predicted from intelligence
  tests)

pointers to other synsets
Type of pointer
POS
Pointer
See wndb
7
Pointer symbols

• For nouns
• !    Antonym _at_    Hypernym      Hyponym m  
   Member holonym s    Substance holonym p  
   Part holonym m    Member meronym s  
   Substance meronym p    Part meronym  
   Attribute    Derivationally related form    
     

See wninput
8
WordNet index file
lemma (word)
POS
pointers
pointers

• abomination n 3 2 _at_ 3 0 09613960 07401317
  00734041

synset file offset
synsets
9
WordNet tools

• Many, many tools
• General documentation
• http//wordnet.princeton.edu/doc
• Online query and lookup
• http//wordnet.princeton.edu/perl/webwn
• APIs and tools http//wordnet.princeton.edu/link
  s
• WordNetsimilarity
• http//wn-similarity.sourceforge.net/
• WordNetsimilarity web interface
• http//marimba.d.umn.edu/cgi-bin/similarity/simila
  rity.cgi

10
WordNet and WSD

• Milhalcea 2002 describes system to sense encode
  text using WordNet (and related tools and
  resources)

11
Milhalcea 2002

• Some tools and resources described
• Senseval
• http//www.senseval.org/
• Evalutation exercises for Word Sense
  Disambiguation
• Senseval-1 3, held in last several years,
  workshops at ACL
• Senseval-4 coming up
• Data and materials from Senseval-3 can be
  downloaded
• Some useful materials for multiple languages
• Materials and test data for English, Italian,
  Basque, Catalan, Chinese, Romanian, and Spanish

12
Milhalcea 2002

• Some tools and resources described
• Semcor
• Sense tagged Brown corpus
• Created at Princeton
• Used for training WSD systems
• Can be downloaded from Milhalceas web site
• http//www.cs.unt.edu/rada/downloads.html
• Were also planning on installing it on Pongo

13
McCarthy et al 2004

• Task find the predominant word senses in
  untagged text
• Unlike Milhalcea 2002, did not rely on supervised
  method using SemCor
• Built a thesaurus from raw text and Wordnet
• Intuition word sense more likely to be
  determined from untagged corpus from context,
  affected by genre, domain or text type
• Rather than relying on SemCors 250,000 words,
  where the word senses are rather limited

14
McCarthy et al

• Thesaurus development relies on dependencies
  between neighbors
• Look at distributional similarities between a
  word and its neighbors

15
McCarthy et al

• Experimented with several similarity measures
  available in WordNetsimilarity
• First experiment used SemCor to see how well the
  unsupervised system worked
• 2595 polysemous nouns in SemCor

16
McCarthy et al

• Experiment 2 against SENSEVAL-2 English All
  Words Data
• Comparison between the precision and recall for
  SemCor vs. their automatic data (and the SENSEVAL
  ceiling)

17
McCarthy et al

• Some experiments with domain specific corpora
  gave these results