Cross-lingual RDF Thesauri Interlinking

Cross-lingual RDF Thesauri Interlinking Tatiana Lesnikova, J´erˆome David, J´erˆome Euzenat INRIA & University of Grenoble Alpes, Grenoble, France tatiana.lesnikova@inria.fr, jerome.david@inria.fr, jerome.euzenat@inria.fr Abstract Various lexical resources are being published in RDF. To enhance the usability of these resources, identical resources in different data sets should be linked. If lexical resources are described in different natural languages, then techniques to deal with multilinguality are required for interlinking. In this paper, we evaluate machine translation for interlinking concepts, i.e., generic entities named with a common noun or term. In our previous work, the evaluated method has been applied on named entities. We conduct two experiments involving different thesauri in different languages. The first experiment involves concepts from the TheSoz multilingual thesaurus in three languages: English, French and German. The second experiment involves concepts from the EuroVoc and AGROVOC thesauri in English and Chinese respectively. Our results demonstrate that machine translation can be beneficial for cross-lingual thesauri interlinking independently of a dataset structure. Keywords: Cross-lingual Data Interlinking, owl:sameAs, Thesaurus Alignment 1. Introduction In the Semantic Web, entities are described in triples (sub- ject, predicate, object) following the W3C Resource De- scription Framework (RDF) (Lassila and Swick, 1999). Most commonly, entities are real-world individuals and events. Moreover, linguistic resources such as thesauri and dictionaries are also available in RDF. These linguistic re- sources should be interlinked to enhance their interoper- ability (McCrae et al., 2012). In this paper, we evaluate a translation-based interlinking method on terminology expressed in different natural lan- guages. This interlinking method has been applied to the encyclopedic resources from DBpedia (Bizer et al., 2009) in English and XLore (Wang et al., 2013) in Chinese on which we obtained good results (Lesnikova et al., 2014). Though this method has been initially developed for cross- lingual interlinking RDF instances, we consider its appli- cation to linking heterogeneous multilingual linguistic re- sources. There are many lexical-semantic resources for different languages and domains grouped in the Linguistic Linked Open Data cloud1 (Chiarcos et al., 2011). We address the following problem: Given two thesauri with labels in two different languages, find equivalent concepts and link them using owl:sameAs relation. We represent concepts as text documents containing labels in a respective language, documents are translated and rep- resented in a vector space model. Similarity between docu- ments is taken for similarity between concepts. Extraction of matches is based on the similarity values. A set of RDF statements forms a labeled directed graph where nodes rep- resent concepts and edges represent relationships between these concepts. An RDF thesaurus is a graph where re- sources are concepts labeled in natural languages. A con- text of a concept are the labels of the neighboring nodes. However, a context can be very narrow if there is not much textual information in the concept description. The remainder of the paper is structured as follows. Sec- tion 2 presents related work on multilingual lexical re- 1http://linguistics.okfn.org/resources/llod/ sources and methods for their interlinking. Section 3 briefly presents the approach for interlinking multilingual entities. Section …

Content truncated to 500 words