FRODO Vision and Project Goals

The most important activity of today's enterprises is the successful and efficient processing of information and knowledge. Knowledge about an enterprise's technologies, products and processes, organization, relations to customers and suppliers, and the particular competencies of its employees constitute the most valuable assets of the company. Knowledge management investigates methods and tools to improve the processing of this knowledge on all levels of the enterprise.

A key aspect in such scenarios is the realization of so-called organizational memories (OM) which capture, store, distribute, and update relevant enterprise knowledge. Current research in the KnowMore and VirtualOffice projects has been focusing on the integration of formal and non-formal knowledge sources by the use of suitable formal knowledge descriptions and the transformation of informal into formal knowledge. Information retrieval and document analysis and understanding (DAU) have been extended by background knowledge and integrated with process modeling and workflow techniques in order to effectively provide context-specific information for supporting knowledge-work activities.

In order to further develop the KnowMore OM approach towards a real technology the FRODO project tackles the following research questions reflecting the real-world requirements of today's enterprise environments:

• How can OM systems efficiently be integrated into the constantly changing world of enterprise information infrastructures?

Each company possesses different legacy information systems and requires other OM services. Operational OM solutions must be configurable from standard modules which link together distributed existing systems, and must be easily adaptable and extendible in order to meet changing requirements and to integrate new services. Distributed inferences must enable scalable solutions.

Thus, FRODO will provide a flexible, scalable OM framework for evolutionary growth.
 

• How can ontology construction and evolution be supported by informal to formal information transformation?

In order to cope with the rapid dynamic changes today's enterprises are facing from both inside and outside, OMs must be designed to incorporate new knowledge automatically or semi-automatically as far as possible. Techniques from document analysis and understanding (DAU) will be adapted to facilitate the task of ontology construction and evolution.

Thus, FRODO will provide a comprehensive toolkit for the construction and maintenance of domain ontologies.
 

• How will users get their personalized information from the various input streams of unstructured documents?

An OM typically processes large amounts of incoming informal documents, which have to be analyzed, filed in the OM for later access and distributed to interested users. Thus, the OM has to bridge the gap between documents and users, describing their information needs with personal profiles, by employing DAU techniques. Today, most DAU systems are tailored to specific applications and adaptation is still one of the main problems to overcome. Furthermore DAU systems use their own knowledge base and are only loosely connected with the OM. A better integration of DAU into the OM allows a better usage of domain specific knowledge to produce more accurate results.

Thus, FRODO will improve information delivery by the OM by developing more integrated and easier adaptable DAU techniques.
 

• How can OM systems successfully be introduced in companies?

The integration of OM technology into the work processes of an enterprise is a complex and difficult task which can profit not only from formalized workflows as a basis for proactive user support, but also from the methodology developed for the introduction of business process modeling technology in the enterprise. While this offers a reliable foundation it has to be considered that interesting knowledge-intensive activities in an enterprise are seldom adequately representable by strictly formalized workflows but observe much weaker dependencies and sequence constraints.

Thus, FRODO will develop a methodology and tool for business-process oriented knowledge management relying on the notion of weakly-structured workflows.
 

The output of the FRODO project will be a formal, logic-based OM framework, implemented by an Intranet-enabled agent platform which provides representation, communication, and inference services especially tailored for the realization of distributed OM applications.

The agent platform shall allow an easy incorporation of new and legacy knowledge sources and services by appropriate wrappers and facilitators. Conceptually, systems realized on top of this agent platform will be a further development of the KnowMore three-layer OM architecture, while technically being instances of Gio Wiederholds Wrapper-Facilitator-Mediator architecture for Information Integration systems.

The coherent design will encompass all modules of the application, the knowledge description, and the knowledge source level, and will provide tailored communication primitives and DAU modules for the OM scenario. This will allow for deeper cooperation and synergies between the manifold sources of formally represented and informal knowledge prevalent in business-process descriptions, ontological structures, user profiles, web-based documents, etc.

Our work will promote the general understanding of enterprise information logistics and will amplify the benefits of Wiederholds approach.

Since all transitions between informal representations and formal inferences are both practically important and scientifically interesting, our classical expertise in document analysis and understanding will be revisited and further developed as a key enabling technology for OM systems. Regarding a systematics of DAU application development, the redesign of DAU functionality in the light of the OM agent framework will lead to a deeper understanding of basic DAU building blocks and their play-together in an application context. Regarding DAU functionality, the close cooperation of DAU components with manifold sources of OM background knowledge, as well as the easy exchange of background knowledge chunks, expectations, intermediary results, user-centric particularities, etc. give possibilities for improving result quality and for envisioning self-adaptive capabilities. By the way, the population of the OM agent platform with existing as well as to-be-developed DAU modules yields a proof-of-concept of the platform's wrapping, representation, and cooperation mechanisms.

While informal documents will be a major source of object level knowledge in an OM, the basis for formal inferences are enterprise information and domain ontologies, the construction and evolution of which is still an open research problem. Pursuing and deepening the approach initiated in KnowMore, we will deliver novel ontology maintenance techniques which extend the manual ontology editing approaches of today by exploitation of text and document analysis in the OM object level as well as feedback from the ontology in use.

All these technical developments constituting a formal and implementation basis of OM systems, will be complemented by appropriate methodological support. All technical developments can only be successful on the basis of a thorough domain and business task analysis of an organization. Moreover, this analysis regards in particular the knowledge-intensive processes which are hard to analyze by nature. Further, such an analysis has to be done in compliance with known and existing modeling endeavors already undertaken in the organization. So we will combine and further develop existing modeling approaches from knowledge engineering, agent-based system development, business-process modeling, and object-oriented system analysis to define a coherent methodology and tool for business-process oriented knowledge management.

About this document ...

This document was generated using the LaTeX2HTML translator Version 98.2 beta6 (August 14th, 1998)

Copyright © 1993, 1994, 1995, 1996, Nikos Drakos, Computer Based Learning Unit, University of Leeds.
Copyright © 1997, 1998, Ross Moore, Mathematics Department, Macquarie University, Sydney.

The command line arguments were:
latex2html -split 0 lv2

The translation was initiated by Andreas on 2000-03-28

2000-03-28