COGNITIVE SYSTEMS, Volume 5, issue 3,4
Contents and Abstracts

to journal index...

 

G.J. Dalenoort,
Introduction 197-198

D. Gernert
,
Utilization of complementary ex pert knowledge by computer-assisted model synthesis 199-210

P. Forbrig, E. Schlungbaum
,
Model-based approaches to the development of interactive systems 211-224

A. Dittmar
,
Task-modelling - psychological and formal aspects 225-240

I.D. Craig
,
Cognitive agents, communication and self-models 241-254

J. Sefranek
,
Conceptual interfaces 255-270

C. Eckert
,
Computer support for early design in aesthetic design domains 271-290

G. de Haan
,
The design of an information infrastructure to support system managers and business procedures 291-302

G.J. Dalenoort
,
The role of internal representations for the performance of tasks 303-321




UTILIZATION OF COMPLEMENTARY EXPERT KNOWLEDGE BY COMPUTER-ASSISTED MODEL SYNTHESIS
Dieter Gernert
University of Munich, Germany

In the construction of descriptive models or decision models for socio-economic or socio-technical systems interdisciplinary cooperation is indispensable. Every contributor has a specific education, training, worldview, etc.; hence models stemming from different authors have their specific advantages and deficiencies. This situation is characterized as “complementary expert knowledge”. Model synthesis is the generation of a model on the basis of several models which already exist or which are built separately by different authors, or on the basis of parts thereof. After an overview of image models in general and of the basic questions of model synthesis several practical approaches are outlined. Without details of an implementation, possible strategies for a computer-assisted model synthesis are presented and illustrated by an example.



MODEL-BASED APPROACHES TO THE DEVELOPMENT OF INTERACTIVE SYSTEMS
Peter Forbrig and Egbert Schlungbaum
University of Rostock, Germany

Model-based development of interactive systems and its supporting tools are an emerging technology. During the last ten years many systems and research prototypes have demonstrated the feasibility of this approach. Up to now the systems are mainly focused on the development of user interfaces only. An overview of some of these systems and their underlying ideas are discussed within this paper.
On the basis of experience gained in developing our TADEUS project we suggest the combination of both approaches. Furthermore, the different models of TADEUS and their relationships are discussed. The discussion is especially focused on the questions of how many kinds of task models are needed.



TASK MODELLING - PSYCHOLOGICAL AND FORMAL ASPECTS
A. Dittmar
University of Rostock, Germany

The design of user-oriented and task-oriented software can be supported by exploiting task analysis and task modelling. There have been several attempts to use formal task models. Most of them are restricted to the problem of human-computer interaction. A formal task model called “system of nested task hierarchies”, presented in this paper, has been developed on a psychological basis. In this approach we embed an individual task into a wider social context, e.g. the working environment of a company, and take note of task-based cooperation between users. A special relation is introduced to describe the cooperation. The dynamic character of the task model is emphasized.




COGNITIVE AGENTS, COMMUNICATION AND SELF-MODELS
Iain D. Craig
University of Warwick, UK

This paper argues that agents in a multi-agent systems should contain models of themselves and of the other agents (including human agents) with which they interact. The models are dynamic and based upon interaction. We suggest ways in which a Symbolic Interactionist position helps us understand the processes by which these models are maintained and used in agents. In particular, interaction and the taking of alternative roles by agents of all kinds is explored as a fundamental mechanism for the exploitation and modification of these models.



CONCEPTUAL INTERFACES
Ján Sefránek
Comenius University, Bratislava, Slovakia

This paper is devoted to an aspect of man-machine interaction (MMI), called conceptual interface in the sequel. By a conceptual interface we understand a set of concepts used in the communication between man and machine together with the relationships among the concepts.
We discuss the basic features of conceptual interfaces. The discussion is based on some examples. The impact of a conceptual interface on context-dependent and dynamic aspects of man-machine interaction is emphasized.
Special attention is devoted to a kind of conceptual interface - Structured Maps: (SM). A structured map defines a database schema and its instance over an information universe. Therefore, an additional structure is superimposed over a world (of documents) and the structure extends the possibilities of querying and searching.
We generalize the concept of SM. We will consider a Society of Structured Maps (SSM, a set of alternative SMs, they represent a set of alternative views on the information universe). Additionally, a set of transformations between the Structured Maps is considered. The transformations provide a dynamic view of the information universe. Finally, it is outlined how a society of SMs can be formalized in terms of Dynamic Kripke Structures.



COMPUTER SUPPORT FOR EARLY DESIGN IN AESTHETIC DESIGN DOMAINS
Claudia Eckert
Deptartment of Design and Innovation, The Open University, Milton Keynes, UK


The early stages of aesthetically driven design processes, for consumer products or clothes, is open to scientific investigation like design in other domains. Designers in visuospatial aesthetic domains have strong mental imagery skills and employ a lot of tacit knowledge. This paper reports on the findings of an extensive observational and experimental study of the knitwear design process, a domain where technical and aesthetic issues are closely linked. In knitwear design sources of inspiration drive idea generation by setting a context for a collection. Most new designs are generated by adapting sources of inspiration following a number of clear strategies. The idea generation process can be supported by computers in two ways: by given designers access to a large number of sources of inspiration, and helping them to explore design spaces through generative methods. The creation of individual designs can be supported by assisting the adaptation of sources of inspiration and by automatically completing partial designs.



THE DESIGN OF AN INFORMATION INFRASTRUCTURE TO SUPPORT SYSTEM MANAGERS AND BUSINESS PROCEDURES
How to Catch a Guru with Quality

Geert de Haan
University of Technology, Eindhoven


This paper describes the information-requirements analysis as part of a project to improve the information-support for a systems management department of an international company which provides software services to its customers according to a variety of formalized processes, procedures, and standards.
Following an extensive analysis of these processes, observations, and semi-structured interviews were held with group managers and system specialists. The following observations were made and will be discussed:



THE ROLE OF INTERNAL REPRESENTATIONS FOR THE PERFORMANCE OF TASKS
G.J. Dalenoort
University of Groningen, The Netherlands

In the attempt to understand the role of human thinking on the performance of tasks, it is necessary to make models of the organization of our knowledge, and the way we use that knowledge in thinking. Two basic types of models can be distinguished: the ones based on the current digital computer, and the ones based on the metaphor of the brain, as far as we understand it: network models. Although formally both types of models may be considered equivalent, there are very concrete differences with respect to implementation and biological plausibility of the actual processes. A number of aspects of network-type of models are analysed, notably the problem of binding, which is relevant for the question how procedural knowledge can be applied in actual situations. It is argued that in the computer-type of models this problem often remains hidden. The use of hybrid systems for understanding chracteristics of human thinking and occurrence of typical errors, has the risk of muddling the fundamental issues involved, and may only be used after careful analysis of the implicit assumptions being made.


 


to journal index...



ESSCS european society for the study of cognitive systems