Information Extraction: New Developments in Astronomical Information Retrieval for Electronic Publications

The explosion of on-line services and the rapid evolution in information technology, with the advent of the WWW, gives its full dimension to the electronic publication. Electronic publication has to be conceived with links to external resources (databases, bibliographic ser- vices) and with intelligent information retrieval tools. To provide links one needs to recognize the relevant information from a document, and to connect this information to the proper distributed resource. Recog- nition is the rst step; the whole procedure may include the validation for correctness and completeness, and the addition of dynamic links to distributed services. In addition, publication in electronic form permits new methods to access published information. Several activities take place at CDS in this context: CDS develops and maintains links to and from other distributed ser- vices (CDS services with electronic publications and the ADS); CDS develops and maintains services which give access to published information (the VizieR catalogue browser for published tables or SIM- BAD which tracks object citations in papers); CDS develops information retrieval tools (the bibliographic maps or tools to automatically recognize object names in a text). All these developments require close connections with the distributed ser- vices (editors, database managers, service managers, ...). A few examples will be presented.