WERKL: A distributed intelligent multimedia information system architecture

Problem Statement

In the course of this project [Herzog et al.1995, Herzog&Petta 1995, Herzog&Petta 1996], we defined the conceptual framework of an architecture for intelligent multimedia information systems, which specifically addresses the problem of providing adequate support to users working in open-world domains. In contrast to closed-world settings which are characterized by a clearly delimited subject matter and the availability of exhaustive and adequate conceptualizations for the tasks to be undertaken, activities in open-world domains show higher variability and routinely require access to information from an unpredictable range of areas in which users may have little prior expertise. Typical examples include high-level decision making; design tasks; or work in areas of rapidly changing contents (e.g., requiring access to information about the latest range of products or recent technological or scientific material). The proposed research is aimed at investigating a novel means of providing computer-based support to these activities.

The solution approach: an agent-based architecture for exploration

We introduced the notion of exploration of the information space as guiding metaphor for our planned research activities in this scenario, resting on the assumption that a large proportion of information needs occurring in open-world domains can be more adequately satisfied by an explorative rather than a search approach. This exploration process integrates human and computer activities into an iterative framework, where subject matters that possibly were previously unknown to the user and related information sources of relevance for the task at hand are identified and subsequently examined. The assessment of the content of subject areas will usually be based on established conceptualizations, but could also be carried out by inspecting extensional representations of the domain (e.g., as captured in collections of multimedia documents) which in turn may lead to the construction novel interpretations, i.e., the definition of a new vocabulary capturing relationships of relevance to the user in the present work context. As enforced by the distinguishing property of open-world settings, namely the preclusion of a complete formalization, exploration cycles have to be carried out in the context of a continuous human-computer collaboration --- as opposed to the traditional delegation of single closed-world tasks to the machine. Consequently, the proposed research will have to consider the integration of knowledge about the accessible subject areas, about the given information need, and about the users themselves.

In continuation of our work we want to build on this existing conceptual basis, and exploit agent-based technology at different levels to progress towards a specification of a complete architecture which is to be applied in an empirical investigation. Theoretical work will place an emphasis on the three adjacent and interdependent areas of representing the content of open-world domains and sharing of this information (e.g., ontological engineering, agent communication language, agent-based interoperability); methods supporting the iterative exploration process, e.g.\ capable of identifying relevant information sources (``passive'' matchmaking and brokering, ``active'' and ``pro-active'' indexing and reconditioning of the accessible and dynamically changing information, etc.); and high-level human-computer collaboration including strategies for graceful recovery from breakdown situations (interface agents). The accompanying applied work is to comprise the design of a prototype of an agent-oriented system, that will allow for an empirical assessment of the researched topics in one or more selected demonstration domains.

A topical application domain: Education

Our extended involvement in educational activities at different levels --- ranging from development of online courses for selected university lectures (e.g., Chemische Übungen, PKU, Hypothesengeleitete Diagnostik) [Petta et al. 1996,Botz et al. 1997, Gahleitner et al. 1997] to ongoing reforming activities concerning the computer science curriculum at the Vienna University of Technology and the curriculum of the Medical School of the University of Vienna --- have provided us with access to resources (human contacts as well as materials) that form an apt and topical application area. As epitomized by recently started efforts such as Educom's Instructional Managemen Systems project or the general move towards integrative and ``open'' problem-based learning methodologies, an environment has originated which invites the design and application of a system facilitating the locating and integration of dispersed pieces of static information and dynamic information-handling procedures.

Publications:

Botz A., Petta P., Haider D., Rapf K., Kremser K., Maerz R.:

Erfahrungen mit einem selbst entwickelten, WWW-basierten Lernprogramm für das Medizinisch-Chemische Praktikum, Zeitschrift für Hochschuldidaktik, Österreichische Gesellschaft fuer Hochschuldidaktik, 20(4), 1997.

Gahleitner-Schick S., Kettenbach J., Gahleitner A., Petta P., Veitl M., Heinz G., Imhof H.:

"Vienna-Blend"; An interactive multimedia tutorial database for MRI, European Congress of Radiology, ECR 97, Vienna, Austria, 3-7 February, Informatique, Suppl. to Vol.7 European Radiology 1997, 1997.

Herzog M., Petta P., Kuehn C.:

Retrieval as Exploration of Large Multimedia Document Bases, in: Working Notes of the IJCAI-95 Workshop on Intelligent Multimedia Information Retrieval, 14th International Joint Conference on Artificial Intelligence (IJCAI-95), Montreal, 131-143, 1995.

Herzog M., Petta P.:

Multimedia Information Systems in Open-World Domains, in: First International Workshop on Multimedia Information Systems, Doubletree Hotel National Airport, Arlington, VA, September 28-30, Department of Computer Science, University of Maryland, College Park, MD, 13-16, 1995.

Herzog M. Petta P.:

Knowledge Representation in WERKL, an Architecture for Intelligent Multimedia Information Systems, in Vouros G.A.(ed.), Proceedings of the ECAI'96 Workshop on Knowledge Representation for Interactive Multimedia Systems: Research and Experience (KRIMS'96), ECAI'96, Budapest, Workshop Notes, August 12, 17-25, 1996.

Petta P., Kettenbach J., Gahleitner A.:

The Interactive Computer-Assisted MR Instruction Program: "The Vienna Blend", Mednet&Datamedica Rio 96, Colégio Brasileiro de Cirurgiões, Rio de Janeiro, May, 1996.