Service Communautaire d'Information sur la Recherche et le Développement - CORDIS

Conceptual retrieval of information using semantic dictionary in three languages

Increased availability of on-line text (due to inexpensive mass storage, accurate optical character readers and high-speed networks) leads simultaneously to a demand for an easy and precise access to the multilingual textual databases without: learning any artificial language; knowing textual database semantics, mastering the language of the document's author. The CRISTAL project concerns the designing of a concept-based information retrieval system. Such a system provides access to textual information of newspapers articles by matching the query and the text concepts, rather than by matching words to words. To achieve such a purpose, CRISTAL involves the integration of linguistic methods and information retrieval techniques. Current research is carried out along two development axes: the adaptation of the Conceptual Dictionary for Multilinguality, the development of a concept-based information retrieval environment, which includes an indexing module, a matching module, a dialogue management module and a multilingual 'free' language query interface. The prototype is developed on top of an existing document retrieval system. Results have already been obtained and the following elements are available:
the software tool to manage, handle, consult and update the Multilingual Conceptual Dictionary;
the analyser processing the query for Italian, French and English;
the algorithm to resolve the ambiguity between word senses, based on the topology of the conceptual dictionary, public deliverables reports for the application requirements and for the re-usable components and tools.

Reported by

Cap Gemini Innovation
86/90 rue Thiers
92513 Boulogne-Billancourt
France
See on map