Arabization and multilingual knowledge sharing

Abstract

The Work Package 2 (WP2) “Networking and Research Setup” aims to enable close and sustainable scientific cooperation between EU scientists and BZU Sina Institute and to enhance the institute’s capacity in scientific cooperation through carrying out several activities: - Facilitate co-supervision of PhD students (Task 2.1), - Set up joint research and cooperation (Task 2.2), and - Encourage co-authoring of scientific articles (Task 2.3), - Self evaluation (Task 2.4). This Deliverable is the final report on the consortium's activities carried out in the Task 2.2: “Set up joint research and cooperation”. It covers all activities and outcomes of research and cooperation setup, which can be summarized as the following: Four EU multilingual knowledge portals developed in previous EU FP7 and eTen projects have been Arabized, i.e., extended to support Arabic: MICHAEL, in section 3, is a thesaurus-oriented portal for the promotion and valorization of digital cultural content. This cultural heritage portal provides a multilingual service to explore digital collections from museums, archives, libraries, etc. Unfortunately, MICHAEL was not full Arabized, as the source code of MICHAEL portal was not available during the project, hence only some related tasks were carried out, such us partial translation of MICHAEL’s thesaurus. The project board decided to Arabize the two other portals: OKKAM and Organic.Edunet. KYOTO, in section 4, is an ontology-oriented wiki-portal in ecology and environment domain. Unlike other portals, KYOTO supports SPARQL queries over structured data and using a rich ontology. Thus, KYOTO ontology has been mapped to the top and core concepts of Arabic Ontology, which is the important step toward integration into the KYOTO architecture, and the potential to facilitate integration with large scale knowledge bases such as DBPedia and GeoNames. OKKAM, in section 5, is an entity management system, and a spin off by UNITN. OKKAM was Arabized (its Interface, Ontology, content, an searching functionality) to facilitate the disambiguation and linkage of Arabic entities with different languages. The entities that were considered include people, organizations, products, places, and events. Such entities have different names (lexical labels) in different languages, which is a major challenge faced when integrating knowledge from different sources, cultures, and languages. SIERA partners have defined a set of activities in task that aim at resolving and linking Arabic entities with entities in multilingual portals to leverage the knowledge across such multilingualism and culture diversities that the portals support.