MINDS - Project Background

MINDS is based on a variety of technologies developed at CRL in previous projects. These will be integrated in the system to provide a working prototype of data gathering and analysis in multiple languages.

These technologies are -

• Document Management - CRL's machine translation system and extraction systems are based around the Tipster Architeccture document manager.
• Machine Translation - in particular rapid deployment MT, as found in the Corelli and Expedition projects.
• Ontology based natural language processing, using the MikroKosmos ontology.
• Information Extraction - multi-lingual name recognition and information extraction. A new approach using ontologically based information extraction will be developed as part of MINDS.
• Onomasticon - one vital new resource being developed as a part of MINDS and also in CRL's MT projects is a cross language data-base of proper names.
• Automatic language and code set recognition.
• Summarization - multiple methods of summarization are being explored and developed as part of the MINDs project.
• Information Retrieval - simple name based retrieval will be used to produce the multi-lingual document collections used by the MINDS system.
• User Interface Design - CRL has extensive experience in the production of prototype demonstration systems. The MINDS systems will be delivered over the web.