Multilingual semantic database

The EC-funded EuroWordNet (EWN) project aims to build a multilingual database coding basic semantic relations between words for some European Languages (Dutch, Italian, Spanish and English) resembling Princeton WordNet 1.5. In addition to the relations in WordNet 1.5, EWN includes domain labels, cross-category and crosslanguage relations, which are directly useful for multilingual information retrieval. The multilingual structure of the database is given by an InterLingual Index that can be used for language-independent indexing of documents and queries in an information retrieval system. We explore a possible application to cross-language text retrieval based on such conceptual indexing, as an interesting alternative to knowledge-based query expansions and corpus-based approaches such as Cross-Language Latent Semantic Indexing.

Download PDF