The main objective of IBROW is to develop an intelligent brokering service able to retrieve knowledge components from distributed digital libraries, according to stated user requirements. The services will go beyond simple component retrieval and will include dynamic configuration of distributed, heterogeneous applications out of pre-existing components retrieved from different libraries. The components concerned are problem solving methods (generic algorithms) and ontologies. This service will provide software-controlled access to a wide range of distributed and heterogeneous digital libraries of reusable knowledge components, at a level, which abstracts from the underlying technology. In the envisaged scenario digital libraries are viewed as active, competence-based components that encapsulate reasoning services, such as configurable information filters, automatic classifiers and design problem solvers.
The following objectives are important for IBROW:
1. The configuration of intelligent services out of reusable components
2. A language for the characterisation of components (meta-data)
3. Annotation of components with meta-data (manually or semi-automatically)
4. Interoperability of heterogeneous components
5. Compatibility with web standards.
DESCRIPTION OF WORK
In order to achieve the overall goal of IBROW, several subgoals need to be attained
1. Development of IBROW-compatible libraries. In the course of the project, four libraries for different areas will be developed. These libraries will contain implemented components annotated with an XML tag set for knowledge components
2. Development of tools to support the annotation process of knowledge components with UPML. We will design a DTD/XML tag set for UPML with the aim of writing a proposal for a W3C standard for the annotation of knowledge components. The spin-off application "easy access to molecular biology algorithms" contributes to this objective
3. Development of a framework for matching component capabilities to user requirements, and for adaptation and integration of generic knowledge components. We will investigate three techniques for matching. Concerning adaptation and integration of components, the framework should provide knowledge users intelligent support for carrying out the complex operations, while automating the simple ones. Contribution comes from the spin-off "Web workbench for Intranet"
4. Development of Internet services. Two different intelligent reasoning services will be implemented and will be accessible through the Internet. Two spin-off applications contribute to this objective selection of manufacturing technology and configuration of customisable filtering agents on the World Wide Web
5. Tested and validated annotation language for characterising generic knowledge components. This language can be used by third-party library providers for annotating the knowledge components that they want to make accessible to others. The users of the language will be supported by dedicated tools, which will abstract the library annotation process from the specific UPML syntax
6. An infrastructure for interoperability of knowledge components distributed throughout the global information infrastructure
7. A collection of generic and speci fic intelligent reasoning services that can be accessed from everywhere by means of a standard web browser
8. An Intranet service where IT workers can browse, select, modify and configure customisable problem solvers and ontologies for their enterprise-specific problems
9. A methodology to construct specialised software brokers for specific types of problems or tasks (e.g. manufacturing technology classification)
10. A system to support easy retrieval and application of standard algorithms in molecular biology
11. A system allowing users to configure and customise precise filtering agents for information retrieval on the web or in an Intranet.
Funding SchemeCSC - Cost-sharing contracts
08190 Sant Cugat Del Valles (Barcelona)
CA 94305 Stanford
MK7 6AA Milton Keynes
1081 HV Amsterdam