Retour vers liste

Détail de la contribution

Auteur: Andrea SCHALLEY

Co-Auteur(s): Alexander BORKOWSKI, Griffith University, Australia

What would help YOU in your work? - Accessing linguistic resources in a highly targeted way

Abstract/Résumé: In recent years, the need for – and number of – linguistic resources has increased substantially. While many such resources are freely available online (e.g. WALS, Ethnologue), they are generally not searchable outside of restricted predefined parameters, and querying across different resources is virtually impossible. Linguists thus still face the task of having to painstakingly assemble pieces of information to support their research, despite the fact that relevant information might already be captured in electronic resources. They might also miss invaluable information, as they might not know of its existence at all. This poster shows an alternative way forward which is expected to allow researchers to more readily find supporting evidence in electronic resources. It is envisaged that researchers will be able to run queries such as “Is there any evidence for Language X marking categories of knowledge sources? Provide relevant examples of this language from across the resources, and list the knowledge source categories” and will receive suitable materials as query results. Semantic Web technologies provide ways of storing data in a linked format, i.e. essentially webs or networks of linked semantic – and hence meaningful – data points (in what is called RDF model, consisting of huge collections of ‘subject-predicate-object expressions’). Given the right querying mechanism, it is possible to search across these complex networks by moving along the relations (‘predicates’) that exist between data points (‘subjects/objects’). Efforts in the Open Linguistics community are currently aiming at converting existing resources into RDF format or creating such resources from scratch. Once these are available and links according to the Linguistic Linked Open Data cloud have been established, the querying can begin. In this poster, we present the underlying mechanism for querying such resources, thereby demonstrating the power that lies behind using Semantic Web technologies. We show how beneficial it is to link linguistic resources in a format that allows semantic cross-searching. We introduce our custom-built, self-contained software module that allows this kind of searching. It can be embedded via well-defined interfaces into other systems and thus be made widely available for linguistics. Furthermore, we show how the use and interchange of queries as well as an improved presentation of query results can be facilitated by embedding queries in report documents that can be shared among researchers and used by a reporting engine to produce files in a variety of formats, e.g. ones suitable for further editing in major word processors or PDF. A software demonstration will allow audiences to test the system for themselves.