Topics dealt with in the Database-group:
Optimizing content top N queries.
Existing database top N optimization techniques do not suffice to optimize the type of top N queries inherent to content based retrieval. Furthermore, Information Retrieval (IR) top N optimization techniques do not behave very well in a database environment due to a nested loop approach. By horizontally fragmenting the statistics tables used for retrieval, we aim to get the best of both IR and database worlds: qualitively good retrieval in an efficient, flexible, and scalable manner.
Data modelling on the WWW.
Today, nearly everyone has access to the www, and is putting information on it. Unfortunately, due to the un-structuredness of the internet it is hard to find specific information when one is looking for it on the web. Search engines available can only rely on information retrieval techniques and most of the time they lack the desired power in query formulation. Modelling the data on the web, in the same way as for databases, should provide us with the necessary basis for enhancing this query formulation. This of course requires special care for dealing with the included multimedia data and the semi-structured aspects of the data on the web. These kinds of problems are studied in the DMW project.
SUMMER - Secure multimedia retrieval.
Querying multimedia content that is located in separate heterogeneous sources is another challenge, not only because of the distribution aspect, but also because of the security aspects involved (access control, copyrights, etc.). Furthermore, the project pursues a convenient OQL-style way of querying that respects MOA's extensibility.
Modelling of database transactions.
Going from data to operations, the next obvious step is to add processes to an object-oriented datamodel in which the order of execution of methods is restricted (life cycles). A life cycle describes the behaviour of one or some objects from a certain perspective; several life cycles together give a complete picture of the objects behaviour. A tool is under development to check these life cycles. Adding processes to an OO data models may also allow for the modelling of workflow management.
Theory for object algebras in databases.
One of the main themes of our group is to rethink the DBMS architecture so that it can be effectively used in non-standard domains. The object algebra Moa is the still evolving product of our research. The goal of this project is to construct a mathematical theory of object algebras in databases with the aim to generalise the applicability and to improve the realisation of Moa in particular.
Distributed XML Query Processing.
Currently imposed on DB researchers is the quest to have data represented as XML--hierarchical and semistructured document-- while keeping all the robust features of the existing relational DB systems. A part of broader goal of integrating XML into DBMS, this project aims at finding efficient techniques for querying distributed (remote) XML data.
Usage of multimedia for tele-education (finished in 2001)
Multimedia is becoming more and more popular. Its value is also recognised for educational reasons, especially in a tele-education environment. An environment is proposed to let a teacher select fragments of multimedia objects to compose his own course. |
