More and more users of information systems are confronted with very large heterogeneous databases that contain information on subjects or themes that cut across theses databases. The need for single access points is growing very fast. Networking technologies stimulate this need for single access points On the other hand the management of these databases is becoming more and more complex. The I-MASS project will study the problems caused by these developments by developing agent-based tools and the development of the concept of the Virtual Reference Rooms, by means of which access to heterogeneous objects can be realized.
A vast digitising movement nowadays confronts the heritage institutions, such as museums, libraries and archives. These institutions are looking world wide for methods that can help in realizing single point access and the improvement of information management technologies. The rise of heterogeneous information systems is striking. Two test beds in which multi agent systems will facilitate the access and management of large information systems will be built during the project.
This project contributes to the achievement of trustworthy and reliable single access points, which provide insight in the coherence between information stored in distributed and heterogeneous data repositories. The i-MASS project will address the problems mentioned above by:
1. Developing prototypes for new mapping and switching mechanisms to begin establishing (syntactic and semantic) interoperability between sets of heterogeneous meta-data related to digital collections stored in various different databases;
2. Developing a prototype tool-set and technologies to build Multi-Agent Systems (MAS) that facilitate these two forms of interoperability;
3. Developing new dynamic user models that enable intelligent agents to become (more) user responsive.
The overall innovation of the i-MASS project results from the concept of creating agent-based Virtual Reference Rooms for single access to distributed heterogeneous data repositories. The Virtual Reference Room will combine various building blocks based on Multi Agent Systems (MAS). A query from a user will be dealt with by MAS (systems) that perform a search task in different sets of meta-data related to databases containing digital collections of text, images and animations. However, before the agents perform their search task, they will "contextualise" and enrich the query by using new mapping and switching (to be developed) mechanisms as well as (existing) thesauri and reference works. The mapping and switching mechanisms help to solve the problem of (lack of) interoperability between different meta-data standards.
Loaded with this knowledge the agents will perform their search task in the meta-data databases. The feedback will subsequently be tailored to the (individual) user needs through dynamic user modelling and contextualising the results.
I-MASS will run for 33 months. The consortium contains of research organizations, a multimedia publisher, a software house and heritage organizations (i.e. museums, libraries and archives). The last ones will provide for the content of the test beds.
The work will exist of 4 phases:
1. Inventory and definition
This phase is to determine the general requirements for MAS models. For this an overview of the different digital objects, and the used standards for their storage, description and indexing within the heritage institutions is needed. This inventory serves as a basis for the definition of the rules for generating and updating metadata.
In the research phase research will be done on the subjects MAS modelling and development; problems of syntactical and semantic interoperability of content and on user modelling and interaction.
The research results will form input for the development of two test beds to be built and tested in the next phase.
The development phase will consist of the building of two test beds, one to be built in the Limburg (Belgium and the Netherlands) environment and the other one will be situated in Italy. For each test bed MAS systems will be set up.
4. Integration and testing
The MAS systems will be tested in the both test beds for their re-usability. The prototypes will be tested on their technological merits, the integration of content of a diverse character and on their user friendliness.
M1: Detailed R&D objectives, evaluation and validation framework
M2: Basic tool set, architectural framework for MAS modelling, blueprint MAS I (syntactic interoperability), Dissemination and use plan
M3: Test results MAS I (syntactic interoperability) for test bed 1
M4: Results evaluation and validation MAS I (test bed 1), parameters for re-usability concept MAS I in test bed 2, blueprint MAS II (semantic interoperability)
M5: MIDTERM REVIEW, complete evaluation and validation MAS I (syntactic interoperability), test results MAS II (semantic interoperability) for test bed 1
M6: Results evaluation and validation MAS II (test bed 1), parameters for re-usability concept MAS II in test bed 2, blueprint MAS III (user interaction)
M7: Complete evaluation and validation MAS II (semantic interoperability), test results MAS III (user interaction) for test bed 2
M8: Results evaluation & validation MAS III (test bed 2), parameters for re-usability concept MAS III in test bed 1
M9: Complete evaluation & validation MAS III (user interaction), Technology Implementation Plan, final analysis of project results and benefits.
Funding SchemeCSC - Cost-sharing contracts
6200 AW Maastricht
6040 AC Roermond
5911 BX Venlo
6200 MD Maastricht