This is a survey paper that argues that the vector space mode is a natural consequence of the distributional hypothesis (DH). DH states that words that occur in similar contexts tend to have similar meanings. This can be used not only to identify synonyms, but a range of analogies as well (eg., 'water' "flows in a" 'river', 'data "flows in a" 'pipe'). Some of these analogies can become complex, involving numerous terms. Consider the possible mappings of terms between 'the solar system' and 'the Bohr atom model'. What about logic? We turn to Dominic Widdows (2004) book Geometry and Meaning were he analyzes relations like 'bass' "and not" 'fish'. What's the other way of doing this sort of thing? Well, there's Wordnet, in which "nouns, verbs, adjectives and adverbs are organized into synonym sets, each representing one underlying lexical concept."
Today: 0 Total: 1107 [Direct link] [Share]
