Two Is Bigger (and Better) Than One: the Wikipedia Bitaxonomy Project

Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, pages 945–955, Baltimore, Maryland, USA, June 23-25 2014. c⃝2014 Association for Computational Linguistics Two Is Bigger (and Better) Than One: the Wikipedia Bitaxonomy Project Tiziano Flati, Daniele Vannella, Tommaso Pasini and Roberto Navigli Dipartimento di Informatica Sapienza Universit`a di Roma {flati,vannella,navigli}@di.uniroma1.it p.tommaso@gmail.com Abstract We present WiBi, an approach to the automatic creation of a bitaxonomy for Wikipedia, that is, an integrated taxon- omy of Wikipage pages and categories. We leverage the information available in either one of the taxonomies to reinforce the creation of the other taxonomy. Our experiments show higher quality and cov- erage than state-of-the-art resources like DBpedia, YAGO, MENTA, WikiNet and WikiTaxonomy. WiBi is available at http://wibitaxonomy.org. 1 Introduction Knowledge has unquestionably become a key component of current intelligent systems in many fields of Artificial Intelligence. The creation and use of machine-readable knowledge has not only entailed researchers (Mitchell, 2005; Mirkin et al., 2009; Poon et al., 2010) developing huge, broad- coverage knowledge bases (Hovy et al., 2013; Suchanek and Weikum, 2013), but it has also hit big industry players such as Google (Singhal, 2012) and IBM (Ferrucci, 2012), which are mov- ing fast towards large-scale knowledge-oriented systems. The creation of very large knowledge bases has been made possible by the availability of collaboratively-curated online resources such as Wikipedia and Wiktionary. These resources are increasingly becoming enriched with new con- tent in many languages and, although they are only partially structured, they provide a great deal of valuable knowledge which can be harvested and transformed into structured form (Medelyan et al., 2009; Hovy et al., 2013). Prominent examples include DBpedia (Bizer et al., 2009), BabelNet (Navigli and Ponzetto, 2012), YAGO (Hoffart et al., 2013) and WikiNet (Nastase and Strube, 2013). The types of semantic relation in these resources range from domain-specific, as in Freebase (Bollacker et al., 2008), to unspec- ified relations, as in BabelNet. However, un- like the case with smaller manually-curated re- sources such as WordNet (Fellbaum, 1998), in many large automatically-created resources the taxonomical information is either missing, mixed across resources, e.g., linking Wikipedia cate- gories to WordNet synsets as in YAGO, or coarse- grained, as in DBpedia whose hypernyms link to a small upper taxonomy. Current approaches in the literature have mostly focused on the extraction of taxonomies from the network of Wikipedia categories. WikiTaxonomy (Ponzetto and Strube, 2007), the first approach of this kind, is based on the use of heuristics to determine whether is-a relations hold between a category and its subcategories. Subsequent ap- proaches have also exploited heuristics, but have extended them to any kind of semantic relation expressed in the category names (Nastase and Strube, 2013). But while the aforementioned at- tempts provide structure for categories that sup- ply meta-information for Wikipedia pages, sur- prisingly little attention has been paid to the ac- quisition of a full-fledged taxonomy for Wikipedia pages themselves. For instance, Ruiz-Casado et al. (2005) provide a general vector-based method which, however, is incapable …

Content truncated to 500 words