BabelNet meets Lexicography: the case of an automatically-built multilingual encyclopedic dictionary

By November 17, 2016,
Page389-397
AuthorRoberto Navigli
TitleBabelNet meets Lexicography: the case of an automatically-built multilingual encyclopedic dictionary
AbstractIn this paper we provide a first study of the lexicographic quality of BabelNet, a very large automatically-created multilingual encyclopedic dictionary. BabelNet 2.0, available online at http://babelnet.org, covers 50 languages and provides both lexicographic and encyclopedic knowledge for all the open-class parts of speech. It is obtained from the automatic integration of several language resources, namely: WordNet, Open Multilingual WordNet, Wikipedia and OmegaWiki. Here we present a first analysis of the dictionary entries in terms of their coverage of English and Italian word tokens in a large corpus and in comparison to existing, well-established dictionaries, namely the Oxford Dictionary of English and the Treccani Italian dictionary. We observe that BabelNet contains most meanings of the frequent words under analysis and provides additional, often domain-specific meanings and their textual definitions unavailable in traditional dictionaries, as well as encyclopaedic coverage for those words.
SessionLexicography and Language Technologies
KeywordsMultilinguality; Encyclopedic dictionaries; Quality evaluation of automatically-created dictionaries
BibTex
@InProceedings{ELX2014-028,
author={Roberto Navigli},
title={BabelNet meets Lexicography: the case of an automatically-built multilingual encyclopedic dictionary},
pages={389-397},
booktitle={Proceedings of the 16th EURALEX International Congress},
year={2014},
month={jul},
date={15-19},
address={Bolzano, Italy},
editor={Abel, Andrea and Vettori, Chiara and Ralli, Natascia},
publisher={EURAC research},
isbn={978-88-88906-97-3},
}
Download