Anders Ardö and Koraljka Golub, IT, LTH

AI Seminar, April 8th, 15-17, glasburen

Title:  Activities at Knowledge Discovery and Digital Library Research Group
        (KnowLib), Digital Information Systems, Dept. of Information Technology

Abstract:
  The KnowLib research group focuses on knowledge discovery and
  development of distributed knowledge organization technologies and
  information systems (Digital Libraries). In particular we explore the
  usability of knowledge organization systems (thesauri,
  classifications, taxonomies, ontologies, gazetteers) in the digital
  library context. Key areas include Information retrieval,
  distributed computing, automated subject classification, focused Web
  crawling, subject specific search engines, and subject browsing. 

  KnowLib is currently involved in the following projects:

    * KLIC-DDL : Intelligent Components of a Distributed Digital Library,
      funded by Vinnova, Swedish Agency for Innovation Systems. 

      The project is aimed at implementation of information services using
      intelligent components in a distributed digital library.

      The work will concentrate around innovative components interacting
      with the architecture for a distributed digital library. In the
      project we will also study methods for integration of heterogeneous
      information sources covering information retrieval, formats, relevance
      ranking and structuring of results for presentation.

    * DELOS (http://www.delos.info/) Network of Excellence on Digital
      Libraries, funded by the EU's Sixth Framework Programme.

      The DELOS network intends to conduct a joint program of
      activities aimed at integrating and coordinating the ongoing research
      activities of the major European teams working in Digital Library -
      related areas with the goal of developing the next generation Digital
      Library technologies.

    * ALVIS (http://www.alvis.info/) - Superpeer Semantic Search Engine,
      funded by EU's Sixth Framework Programme.

      The ALVIS project conducts research in the design, use
      and interoperability of topic-specific search engines with the goal of
      developing an Open Source prototype of a
      distributed, semantic-based search engine. Distribution is intended to
      be able to work with heterogeneous search servers, using query topics
      as a routing mechanism, and using distributed methods for ranking.

      The approach is not the traditional Semantic Web approach with coded
      meta-data, but rather an engine that can build on content through
      semi-automatic analysis. Linguistic processing is right in the
      heart of the search engine which uses a probabilistic document model for
      relevance evaluation and information retrieval.  Natural language
      processing and statistical processing of the corpus extracts semantic
      information from both queries and documents thus providing
      disambiguation, semantic clarification and topical content.  The
      automatically extracted information can enrich indexed documents and
      can be used to estimate the relevance of documents to queries.

    * User behavior in the Renardus Web service, a study based on log
      analysis.