 | |
| Stable release | BabelNet 5.3 / December 2023 |
|---|---|
| Operating system | |
| Type | |
| License | Attribution-NonCommercial-ShareAlike 3.0 Unported |
| Website | babelnet |
BabelNet is amultilingual lexical-semanticknowledge graph,ontology and encyclopedicdictionary developed at theNLP group of theSapienza University of Rome under the supervision ofRoberto Navigli.[1][2] BabelNet was automatically created by linkingWikipedia to the most popular computationallexicon of theEnglish language,WordNet. The integration is done using an automatic mapping and by filling in lexical gaps in resource-poorlanguages by usingstatistical machine translation. The result is anencyclopedic dictionary that providesconcepts andnamed entitieslexicalized in many languages and connected with large amounts ofsemantic relations. Additional lexicalizations and definitions are added by linking to free-license wordnets, OmegaWiki, the EnglishWiktionary,Wikidata,FrameNet,VerbNet and others. Similarly to WordNet, BabelNet groupswords in different languages into sets ofsynonyms, calledBabelsynsets. For each Babel synset, BabelNet provides short definitions (calledglosses) in many languages harvested from both WordNet and Wikipedia.
As of December 2023[update], BabelNet (version 5.3) covers 600languages. It contains almost 23 million synsets and around 1.7 billionword senses (regardless of their language). Each Babel synset contains 2 synonyms per language, i.e., word senses, on average. The semantic network includes all the lexico-semantic relations from WordNet (hypernymy and hyponymy,meronymy andholonymy,antonymy andsynonymy, etc., totaling around 364,000 relation edges) as well as an underspecified relatedness relation from Wikipedia (totaling around 1.9 billion edges).[1] Version 5.3 also associates around 61 million images with Babel synsets and provides a LemonRDF encoding of the resource,[3] available via aSPARQL endpoint. 2.67 million synsets are assigned domain labels.
BabelNet has been shown to enable multilingualnatural language processing applications. The lexicalizedknowledge available in BabelNet has been shown to obtain state-of-the-art results in:
BabelNet received the META prize 2015 for "groundbreaking work in overcoming language barriers through a multilingual lexicalised semantic network and ontology making use of heterogeneous data sources".
The Artificial Intelligence Journal paper that describes BabelNet[1] won the Prominent Paper Award in 2017.[9]
BabelNet featured prominently in aTime magazine article[10] about the new age of innovative and up-to-date lexical knowledge resources available on the Web.
{{cite web}}: CS1 maint: bot: original URL status unknown (link) Transactions of the Association for Computational Linguistics (TACL), 2, pp. 449-464, 2014.