Department of Software Technology
Vienna University of Technology
Management of Distributed Information
Repositories
Abstract:
The number and size of information repositories available on the Internet is rising exponentially,
requiring more powerful methods for interaction. Apart from mere list-based query interfaces
we face the need for intelligent tools, which are able to learn and utilize user profiles, which can
structure and visualize the information, which can combine and deal with diverse types of
information, be it literature or organizational knowledge assets, and are last but not least
intuitive, i.e. usable without expert knowledge on query languages, logics and the like. Due to
the amount of information available and its speed of growth, these systems need to be highly
scalable and may not require manual interaction to keep them up to date.
The project described in this paper aims at the development of an integrated solution for the
requirements arising from the need to provide strong support at all levels of interaction with a
digital library in the most general way. A flexible architecture based on a neural network model
allows simple scaling as well as application in different and evolving contexts. Furthermore,
`computational intelligence tools' shall assist the user in searching the library, allowing to find
and retrieve stored information by means that go beyond simple keyword search, including user
profiles and retrieval based on sample texts. Strong emphasis is paid to the visualization of the
contents of the information repository allowing intuitive navigation in and understanding of any
collection of information. Standard web-based interfaces shall provide access to distributed
repositories with the possibility to create higher level information pools by combining several
information services on the net based on this architecture.
Up
Comments: rauber@ifs.tuwien.ac.at