WordNet
In plain wordsA hand-made map of English vocabulary in which words sharing a meaning are bundled together and the bundles are linked — recording, say, that a dog is a kind of animal and a wheel is part of a car.
A manually compiled resource of English word meanings in which each sense is represented by a synset — the set of words and phrases that can express it — together with a brief gloss. The synsets form separate networks for each of four parts of speech: noun senses are tied together by is-a links (hypernymy and hyponymy) and part-of links (meronymy), verb senses by relations such as entailment, and adjectives largely through antonym pairs. The nouns form a deep taxonomy rooted in 'entity'. Version 3.0 (2006) holds about 155,000 words in roughly 117,000 synsets. Freely licensed and bundled with toolkits such as NLTK, it became the default sense inventory for word sense disambiguationword sense disambiguation, n.: Working out which meaning of a word is intended in context — whether 'bank' means a riverside or a place that keeps money., a source of path-based word-similarity measures and features for pre-neural NLP, and the scaffold of ImageNetImageNet, prop. n.: A huge collection of more than 14 million labeled photos, sorted into object categories, whose yearly recognition contest from 2010 to 2017 sparked the deep learning boom., whose categories are WordNet noun synsets. Wordnets built on the same design now exist for dozens of languages.
We mapped every label to a WordNet synset so that 'puppy' and 'dog' roll up to the same parent class.
Origin
Begun in 1985 at Princeton University's Cognitive Science Laboratory by the psychologist George A. Miller, originally to test theories of how people organize word meanings in memory; Christiane Fellbaum later co-directed it. It was described in a 1990 special issue of the International Journal of Lexicography and in Fellbaum's 1998 MIT Press volume. Princeton's most recent release is version 3.1 (2011); Open English WordNet, a community-maintained fork, has continued to revise the database.
Usage note
Often used as an ontologyontology, n.: A carefully defined vocabulary for a subject such as medicine, spelling out each kind of thing, how kinds nest inside one another, and which connections between them are allowed., though it encodes word-sense relations rather than formal logic. Critics note that its sense distinctions are finer than annotators can reliably apply, and the person categories that ImageNet inherited from it included slurs and offensive labels, many of which the ImageNet team removed in 2019 amid public criticism.
Related
- ImageNetImageNet, prop. n.: A huge collection of more than 14 million labeled photos, sorted into object categories, whose yearly recognition contest from 2010 to 2017 sparked the deep learning boom.
- word sense disambiguationword sense disambiguation, n.: Working out which meaning of a word is intended in context — whether 'bank' means a riverside or a place that keeps money.
- ontologyontology, n.: A carefully defined vocabulary for a subject such as medicine, spelling out each kind of thing, how kinds nest inside one another, and which connections between them are allowed.
- knowledge graphknowledge graph, n.: A huge web of facts stored as connections between things — 'Paris is the capital of France', 'Marie Curie won a Nobel Prize' — that computers can search, combine and reason over.
- word embeddingword embedding, n.: A list of numbers standing for a word, learned so that words used in similar ways get similar lists. It lets a computer measure that 'cat' is closer to 'dog' than to 'car'.
- Penn TreebankPenn Treebank, prop. n.: A collection of English text, best known for its million words of newspaper articles, in which experts marked each word's grammatical role and drew each sentence's structure as a tree; it trained and tested a generation of language software.
Sources
- Miller, G. A. (1995). WordNet: A Lexical Database for English. Communications of the ACM 38(11), 39–41.
- Fellbaum, C. (Ed.) (1998). WordNet: An Electronic Lexical Database. MIT Press.
Mentioned in
Penn TreebankPenn Treebank, prop. n.: A collection of English text, best known for its million words of newspaper articles, in which experts marked each word's grammatical role and drew each sentence's structure as a tree; it trained and tested a generation of language software., semantic networksemantic network, n.: A way of storing knowledge as a web of concepts joined by labeled arrows, such as 'canary — is a — bird' and 'bird — can — fly', so a program can follow the links., ConceptNetConceptNet, prop. n.: A free, crowd-built network of everyday facts — 'a knife is used for cutting', 'rain makes things wet' — in many languages, which AI systems use to make sense of ordinary situations.