HCI Bibliography Home | HCI Conferences | HCII Archive | Detailed Records | RefWorks | EndNote | Hide Abstracts
HCII Tables of Contents: 89-1a89-1b89-2a89-2b91-1a91-1b91-2a91-2b93-1a93-1b93-1c

Proceedings of the Sixth International Conference on Human-Computer Interaction

Fullname:Proceedings of the Sixth International Conference on Human-Computer Interaction
Editors:Uuichiro Anzai; Katsuhiko Ogawa; Hirohiko Mori
Location:Tokyo, Japan
Dates:1995-Jul-09 to 1995-Jul-14
Publisher:Elsevier Science
Standard No:ISBN 0-444-81795-6 ISSN 0921-2647; hcibib: HCII95
  1. HCII 1995-07-09 Volume II. Human Centered System Design
    1. II.1 Decision Making Support System
    2. II.2 Human Factors in Software Engineering
    3. II.3 Support for Creativity
    4. II.4 Interaction in Database Query
    5. II.5 Vehicle System & Aircraft
    6. II.6 Computer Supported Cooperation in Product Design
    7. II.7 Communication Services
    8. II.8 Learning Environments 1
    9. II.9 Learning Environments 2
    10. II.10 Control System for Special Application
    11. II.11 Cognitive Model in Process Systems
    12. II.12 Interface Design and Evaluation in Process System
    13. II.13 Human Factors in Nuclear Power Plants
    14. II.14 Operation Aids Plant Diagnosis
    15. II.15 Intelligent and Adaptive Systems
    16. II.16 Machine Learning
    17. II.17 Intelligent Interface Design

HCII 1995-07-09 Volume II. Human Centered System Design

II.1 Decision Making Support System

User Participation in the Early Phases of Developing Decision Support for Operators in Advanced Manufacturing Systems BIBA 615-620
  A. Johansson
This paper describes a Tasks Evaluation and Analysis Method (TEAM) seen as a framework for user participation. It focuses on the selection of tasks that need support and on giving the users a positive attitude towards their new tool, thereby increasing the chance for the tool to be accepted and effectively used. TEAM has been used for case studies in advanced manufacturing systems.
Trade-Off Decision Making: Choosing between Alternatives in User Interface Design BIBA 621-626
  S. Howard
User interface design (UID) is a complex task, any account of which is likely to be multi-faceted. This paper focuses on one part of UID, the choice between design alternatives, termed Trade-off Decision Making (ToDM). Firstly, ToDM is defined and exemplified. Secondly, a model of ToDM, based on extensive verbal protocol studies of real world UID, is presented discussed.
The Development of a Decision Support System in Marketing Promotion Systems BIBA 627-632
  Cheng-Wen Neu; Sheue-Ling Hwang
A decision support system (DSS) for unstructured decision problem is presented in this study. The major part of this system is the knowledge base which is created by means of system analysis and expertise, and the working system of the DSS is the promotion system which is a part of marketing system. Moreover, an experiment was conducted with two group (with/without a DSS), and each group was asked to solve some decision problems. The result revealed that the performance of the group with a DSS was significantly better than the group without a DSS.
An Approach on Developing an Advice System for Starting a New Business: The Case of a Supermarket BIBA 633-638
  Yumiko Taguchi; Tsutomu Tabe
This paper gives preliminary information on the methodology for developing an advice system based on a knowledge-based system for planners that want to start new businesses such as supermarkets.

II.2 Human Factors in Software Engineering

Display-Based Skills in a Complex Domain: The Use of External Information Sources in Computer Programming BIBA 641-646
  Simon P. Davies
This paper presents evidence for differences in the nature of programmers' information externalisation strategies. Two experiments are reported which suggest that experts rely much more upon the use of external memory sources in situations where the device they use to construct the program hinders the utilisation of a display. Experts tend to externalise low level information, mainly to aid simulation, whereas novice's develop higher level representations which might be characterised as transformations or re-representations of the program. Moreover, in the case of experts the nature of externalised information appears to depend upon whether they are generating a program or comprehending it. These results provide support for a display-based view of problem solving and address strategic differences in the externalisation of information.
Focal Structures in Program Comprehension: Implications for the Design of Programming Support Tools, Debugging Aids and Tutorial Environments BIBA 647-652
  Simon P. Davies
This paper explores the relationship between knowledge structure and organisation and the development of expertise in programming. A study is reported which provides support for a model of knowledge organisation in programming that stresses the importance of knowledge restructuring processes in the development of expertise. This is contrasted with existing models which have tended to place emphasis on schemata acquisition as the fundamental mode of learning associated with skill development in programming. In particular, this paper emphasizes the interaction between a programmer's knowledge structure and salient syntactic elements contained within the code. It is suggested that previous models of programming have tended to consider one or other of these factors in isolation without examining their interaction. Finally, the implications of this model for theories of debugging and for the design of programming environments are briefly discussed.
Factors Influencing the Classification of Object-Oriented Code: Supporting Program Reuse and Comprehension BIBA 653-658
  Simon P. Davies; David J. Gilmore; Thomas R. G. Green
This paper reports a study of the use of card sorts in the categorisation of fragments of object-oriented programs. We are interested in the way in which programmer's think about code so that we might attempt to provide support for browsing and reuse within object-oriented environments. Hence, we have been exploring the use of knowledge acquisition techniques in order to elicit programmer's knowledge about code. Our results showed that experts tended to focus upon the functional relationships between the code fragments, and that the novice group were much more concerned with objects and inheritance. We discuss these results in terms of claims that have been made about the naturalness of conceiving the world in terms of objects and their relationships.
An Editor for Specifications in a Software Development Support System BIBA 659-664
  Tomoo Ikeda; Katsumi Tanaka; Tetsuro Chino; Yutaka Deguchi
We have developed an editor for specifications, which enables a user to facilitate natural and efficient interactions with the software development support system. In this paper, we describe two functions required in the editor. First, we describe a function for preserving the document structure of the specification, enabling the system to maintain consistency of document structure with the user's edit operations. We have developed a document structure editor to implement this function. Second, we describe a function for preserving the correspondence between a specification document and the result of analysis that enables a user to create documents more efficiently. We propose a method, in which we introduce an identifier that is assigned to each sentence in the document, and the editor calculates minimal structural changes made by the operation as the range for analysis. Then, we implemented an editor for specifications, and we describe the result of experiments on our editor. Our editor enables the user to cut character input by 13% and analysis time by 86%.
Behavioral Factors in Software Development BIBA 665-670
  Elif Demirors; Frank Coyle
Software development is essentially a human activity. As a result, factors affecting the behavior of those involved in the development process are of great importance to the software community. We identify three levels at which behavioral factors associated with software development can be investigated: individual behavior, team behavior and organizational behavior. Analysis of studies on individual and team behavior in software development environments reveals the failure of existing tools in supporting team-based software development. Our survey on team practices among software professionals also confirms this fact. To address this situation we propose a software development environment based on a blackboard architecture that supports multi-agent, opportunistic problem-solving. We expect that this architecture holds promise for the next generation software tools by providing a kernel for team-based software development.
An Interactive Specification Acquisition Method for Software Design BIBA 671-676
  Kouichi Sasaki; Kazuyuki Gotoh; Yasuko Nakayama
This paper describes a specification acquisition method used in PlantBASE, an automatic software design system for plant control. In this method, a plant model is first provided with the knowledge about behaviors and functions of each machine composing a target plant. By referring to this knowledge, PlantBASE recognizes the physical structure of the target plant from the specifications given in the form of tables and diagrams and generates a flowchart of machine operations, thereby reducing the cost of describing the control sequence for the plant.

II.3 Support for Creativity

A Basic Framework of Narrative Generation System as Creative Interface BIBA 679-684
  Takashi Ogata; Koichi Hori; Setsuo Ohsuga
In this paper, we describe a basic framework of the narrative generation system that we have developed and its use as a creative interface tool. Actually, because narrative has the useful characteristics for supporting human creative tasks, for example, the function generates a new meaning from fragmentary information, aesthetic representation function, virtual reality function and the function comprehensively represents various knowledge, we can utilize the narrative generation system as a kind of creative interface tool by realizing in the system these functions. Based on this basic idea, we explain the methodology our general narrative generation system and we describe our approach to the idea of creative interface through an experimental attempt, namely, an application of above narrative generation system to marketing/advertising activities, especially scenario generation of TV commercial message.
A System for Assisting Creative Research Activity BIBA 685-690
  Masanori Sugimoto; Koichi Hori; Setsuo Ohsuga
We have built a system for assisting creative research activity. Its main function is to visualize the semantic relation between viewpoints of information. The system is applied to assisting researchers in the field of science and technology in their creative activities, communication / group work and information retrieval (IR). We have carried out several experiments and confirmed that the system is effective for these applications.
A Card-Handling Tool which Supports Multimedia Data and Resource Accessing on the Internet BIBA 691-696
  Yoshihiro Ohmi; Katsutoshi Nakamura; Naohiko Takeda; Kazuhisa Kawai; Hajime Ohiwa
The multimedia application system based on the card-handling and its extension which makes us be able to access information resources on the Internet are described. We considered the multimedia utility and its suitability for card-handling, and developed a card-handling tool KJ-Editor which supports the multimedia data and Internet resource accessing.
Evaluation of a Thinking Support System from Operational Points of View BIBA 697-702
  Kazuo Misue; Kozo Sugiyama
A thinking support system D-ABDUCTOR is evaluated from operational points of view. A novel, well-designed and rigorous method, called LCRM is developed to design tasks for evaluation experiments. An experiment using LCRM was made by five subjects, and 245 tasks were performed in total. Results of the experiment show efficiency of D-ABDUCTOR.
Toward Augmented Creativity: Intelligent Support for Discovering Latent Relationships among Ideas BIBA 703-708
  Yosuke Kinoe; Hirohiko Mori; Yoshio Hayashi
Augmentation of creativity has recently become an important aspect in the design of computer-supported work. In this paper, we propose a methodology for stimulating analysts to formulate new ideas. In particular, we shed light on a process for discovering previously unknown relationships among existing ideas. To support the methodology, we have developed an experimental system that is intended to provide an intelligent support environment. This system is based on the Genetic Algorithm.
   This methodology was effective for stimulating analysts to expand their perspectives on the central issues. Analysts tried to formulate new ideas by discovering previously unknown combinations of ideas. One unique feature of the methodology is a process whereby analysts and a computer system can develop new ideas collaboratively.

II.4 Interaction in Database Query

Sustaining Interaction in Database Query BIBA 711-716
  R. Inder; J. Stader
Current research on database systems breaks the problem into two halves -- formulating a query and presenting the results. This paper suggests that there is a third "half" which deserves attention in its own right: sustaining the interaction.
Query Assistance: The Query Interface of the CORE Shell System BIBA 717-722
  C. Stephanidis; D. Akoumianakis; D. Grammenos
This paper presents the user interface of the CORE shell system and reports on the effort to build a Query Interface Assistant (QIA) to a formal database query language such as SQL. The CORE shell system addresses the domain of Assistive Technology and aims to provide a useful tool for information seeking actors who may not be particularly aware of the organisation or structure of data residing on existing information sources. To this end, a cooperative QIA to the SQL formal database query language has been developed, which guides the user's query formulation, both conceptually and syntactically. At present, a prototypical implementation of the QIA is used to provide access to a selected number of databases, which hold data relevant to the Assistive Technology domain.
Human Interface for an Ambiguous Image Retrieval System BIBA 723-728
  Masaomi Oda
The conventional image retrieval system has assumed that the retriever has pre-conceived image before retrieval. Therefore, the retriever has great difficulty when he or she has no clear target image. We propose a new image retrieval system where the user can retrieve an ambiguous target image through an interactive process. The system displays candidates according to the image in the user's brain. The usability and efficiency of the system were confirmed by experiments.
Calculating Word Similarity Based on Common Attributes, Total Semantic Distance and Concept Frequency BIBA 729-734
  Jin Cui; Hiroshi Yasuhara
Recently, the technique of Case-Based Reasoning [1,2] has attracted considerable attention in the field of Artificial Intelligence. One of the most important concerns when using Case-Based Reasoning is the identification of a basic similarity between sample data and input data. A calculation of the similarities between words is usually necessary. In previously published work, word distance has been calculated using a word thesaurus [3]. However, such challenges as how to treat words that have more than one meaning or concept, and how to calculate the similarities between two words that belong to two different languages remain largely unsolved. The approach utilizing syntactic features to correct the distortion contained in semantic hierarchies [4] may serve in solving some problems, but it is limited to cases where the syntactic features of words are definite and different. The calculation of the proximity scores by the overlap or intersection of actual examples for two words [5] is also a useful approach because the co-occurrence relation roughly includes various factors related to similarity. However, this approach requires a very large text database and is unwieldy to use in a real application situation, because the entire body of data needs to be searched to answer every query.
   In this paper, we will introduce an approach to calculate the similarities between two words based on the concepts of the words; each word concept is treated not only as one of word sense, but also as an element of a concept set. Our method will be introduced in the latter part of this paper with an experiment using the EDR Electronic Dictionary [6].
Development of a GOMS Model of Database Retrieval BIBA 735-740
  H. C. Chan; C. G. Koh; K. K. Wei
Human-computer interaction models are integrated and enhanced to model the process of query writing, where users write queries to retrieve data from databases. A GOMS model detailing goals and operations is developed for the process of query writing using the query language SQL. Through protocol analysis, the model is compared with subjects' actual query retrieval processes.

II.5 Vehicle System & Aircraft

Designing Interfaces for Dynamic System: A Simulated Bus Traffic Task Testing Anticipatory Behavior BIBA 743-748
  S. Mailles; C. Marine; J. M. Cellier
Codes used to design interfaces in bus traffic control tasks may influence anticipatory behavior. Former studies showed that there is a very wide variety of interfaces in work situations. In order to study anticipatory behavior under different coding conditions, a bus traffic simulator was developed. The experiment conducted used four different interfaces. Results found for three different tasks performed with these interfaces were analyzed in terms of reaction time, nature of responses and precision in drawing future situations. They revealed significant differences between the interface types.
Psychological Assessment of Car Navigation System BIBA 749-754
  Kazuaki Namba; Fumio Mizoguchi
Now, Car Navigation System that is becoming widespread in Japan is not so usable. In this study three experiments examined how Car Navigation System support the human cognition or not. We investigated the reason of a person lost his way (experiment 1), whether Car Navigation System and voice guide function had a effect for supporting the human cognition (experiment 2), what was the reason of difficulty of operating Car Navigation System (experiment 3). In this study, we pointed out the reason that a person lost his way was that he had a wrong present position caused by failure of finding a landmark or a node, and he got out the situation by U-turn not by detour, Car Navigation System was quite effective for detecting to lose his way, the interface of Car Navigation System was opened for improvement.
Technology Transfer in Air Traffic Control System: A Brazilian Study of Case BIBA 755-758
  Luiza Helena; Boueri Rebello
Air traffic controlling has always been, and will always be, the key to flight protection. So the goal of this work is to study the implementation of a new air traffic control system in order to improve work conditions. The flight controllers are the main element of decision, even with the complete automation. With the implementation of a new system, there is the impact of the technological change: the controllers have to work with computers all the time.
The Cockpit Assistant System CASSY -- Design and In-Flight Evaluation BIBA 759-764
  Marc Gerlach; Reiner Onken; Thomas Prevot; Wilhelm Ruckdeschel
This paper focuses on the design principles and the realization of an intelligent, knowledge-based, on-board pilot assistant system for aircraft operation under instrument flight rules. The system has undergone a flight test campaign in summer 94. The tests and the results of the campaign are presented.
Cognitive Engineering Approach to the Evaluation of Human-Computer Interaction: The Case of Air Traffic Control BIBA 765-771
  I. Gaillard; P. Amaldi; M. Leroux
Cognitive engineering modeling provides guidelines for designing aiding tools that are "users" driven rather than "technology" driven. Following a cognitive task analysis of the air traffic controller's, an integrated decision support system has been designed. ERATO (En Route Air Traffic Organizer) includes a series of functions to assist controllers during the management of air traffic scenarios. The most important is the information filtering function. It is the basic support of the expert system in that it "highlights" those aircraft that need to be considered when dealing with a selected traffic problem. For a number of predetermined portions of air traffic scenarios controllers were asked to indicate all of the aircraft relevant to the flight of selected airplanes. The differences between ERATO and the controllers' output were taken into account to modify the working principles of the algorithm. Furthermore, the controllers' answers provided further understanding of the knowledge brought into bear to identify traffic problems. Finally, a taxonomy of the knowledge revealed by the controllers is being validated and used to further define the scope of the decision support system.

II.6 Computer Supported Cooperation in Product Design

Simulating Group Work in Mechanical Engineering Design Departments BIBA 775-780
  Friedhelm Steidel
Technical changes like the introduction of CA-techniques are an opportunity to change work organization. However, in general, the introduction of CA-techniques does not go hand in hand with such a restructuring. Empirical case studies in mechanical design departments showed the following problems: the staff was not well qualified to use CAD-systems, CAD was used as an electronic drawing board and the work organization within the design departments had not changed at all. Neither was there a change towards cooperation with other departments like the production planning department [1,2,3]. A lack of efficiency and a suboptimal task distribution to individuals were results of this kind of CAD practice. No positive impact on the length of the run of these projects was achieved.
Imagery and the Design Process. Suggestions for the Development of CAD-Systems BIBA 781-786
  Renate Eisentraut
It is only through understanding how designers think and how they act that user-friendly computer-aided design tools can be developed. If they are linked to research in engineering design, psychological research and its findings on human thought and action thus can contribute to a better understanding of the design process and help improving it.
   Asked to describe their ways of having dealt with an engineering design problem, designers often report the use of mental images. Scientists also emphasize the importance of mental imagery ability for the design process. Frieling & Hilbig [1] as well as Frick & Muller [2] state that, in this point, the invention of new technologies like CAD-systems will not cause any substantial changes.
   This paper purports a characterization of mental images. Secondly, their importance for problem solving and especially for engineering design will be outlined. Finally, suggestions for the development of CAD-systems will be developed.
Introducing a Telecooperative CAD-System -- The Concept of Integrated Organization and Technology Development BIBA 787-792
  M. Rohde; V. Wulf
To cope with dynamics of a changing environment, organizations have to be able to react flexibly. To support organizational change we present the concept of integrated organization and technology development. This approach offers a framework to deal with organizational and technological change jointly in an evolutionary and participative way. We will apply this approach to the introduction of a shared CAD-system which supports users in discussing design problems bridging spatial barriers.
Personal Communication and Telecooperation in Product Design -- Requirements for Telecooperative CAD-Systems BIB 793-798
  J. Springer; D. Herbst; C. Schlick; J. Stahl
Human-Computer Cooperative Problem Solving in R & D by Supporting Users' Changing Viewpoints and by Using Tacit Knowledge BIBA 799-804
  Kazuhiro Ueda; Kiyoshi Niwa; Masashi Shirabe; Sakae Okuda
This paper proposes a human-computer cooperative method of supporting and facilitating expert researchers' problem solving in R & D, which has been scarcely proposed. The proposed method consists of the following two sub-methods; the one is for providing the guides which enable users to create new target concepts by changing users' viewpoints and by making much use of tacit knowledge (CCCV method), and the other is for retrieving concrete methods of realizing the new target concepts from a case-base on the basis of analogical knowledge association by utilizing Multi-Dimensional Scaling (KADS method). The CCCV method is designed based on the analysis of the interview protocols on changing viewpoints in actual scientific problem solving, which approach is important for constructing a useful supporting method in R & D.

II.7 Communication Services

Advanced Services Experiments on a Broadband Communication Network: The Tuscany MAN BIBA 807-811
  V. Cappellini
A broadband communication network (Tuscany MAN) is described. Advanced services experiments are presented regarding telemedicine, teleenvironment, teleeducation, remote access to Museum and communication aids for disabled people.
JULIA: An Intelligent System Allowing Local and Remote Access for Information Requests into Office Communication Terminals BIBA 813-818
  V. Cappellini; L. Lastrucci; A. Mecocci; A. Raggioli
Julia is an intelligent system to reduce user frustration and improve personal communications. It allows an human machine interaction by voice while the speech understanding system is based on a set of keywords. The system is able to draw inference from a known domain and execute some reasoning and answering tasks in response to the user requests. A large collection of sentences is used to address questions to the user when the reasoning process cannot go on. The sentences are selected in order to obtain a natural, not repetitive dialog between the user and the machine.
   An intelligent graphical interface learns the user skill degree and matches its help level to the user needs.
HCI Methods and Tools in Computer-Supported Interpersonal Communication: Towards Error-Free Information Exchange in Human-to-Human Interaction BIBA 819-824
  L. Balint
Computer-supported human-to-human interaction as a means of advanced interpersonal communication is investigated. General aspects and basic concepts are outlined and a straightforward method of using specific computerized tools in the support of human-to-human communication is suggested. The primary goal of the contribution is to motivate and inspire future research into the challenging area of how machines may aid in one of the most subjective human activities, interpersonal communication.
   It is shown that the introduction of specific and intentional computer-based processing of human messages highly enhances the probability of error-free and adequate human-to-human message transfer. By introducing an intelligent machine into the interpersonal discourse, distortion of the factual content in the messages is eliminated or at least drastically reduced.

II.8 Learning Environments 1

Analogy-Based Learning Support Environment for Novice BIBA 827-832
  Yoshio Nakatani; Toyoo Fukuda
Technology makes remarkable progress and people are asked to learn their unfamiliar domain knowledge in a short period. We propose a computer environment which supports learning processes of novice workers who try to learn their unfamiliar domain knowledge, especially abstract knowledge systems, such as computer operating (OS) systems.
   In this paper, we first review the psychological studies about the learning process of the novices in section 2. We propose a learning support environment which considers the learning process in section 3. In section 4, our analogy-based learning environment is explained. We show the software architecture of our system which realizes our framework on the engineering workstation in section 5.
A Learning Environment for Knowledge-Structuring in History BIBA 833-838
  Tomoya Horiguchi; Tsukasa Hirashima; Akihiro Kashihara; Jun'ichi Toyoda
In this paper, we present a system which explicitly helps a student structurize his/her knowledge in history, a subject in most cases learned by rote. In this system, we use two strategies to help the student: Knowledge-Structuring by Visualization (KSV) and Knowledge-Structuring by Reasoning (KSR). The former enables a student to communicate with the system smoothly by using a visualization tool. The latter enables him/her to firmly grasp knowledge by acquiring it through his/her own reasoning. We have implemented our system on a Sun SPARCstation10, and asked ten subjects to learn history with the system. With the result that seven of them reporting that the system contributed to a deeper understanding of history on his/ her part.
How to Make Fill-in-Blank Program Problems BIBA 839-844
  Masato Soga; Akihiro Kashihara; Jun'ichi Toyoda
We expect that students can reinforce their understanding and that they can stabilize knowledge of algorithm, by solving fill-in-blank program problems. In this paper, we propose how to make fill-in-blank problems. We confirm that the relation between blanks affects student's cognitive load heaviness by an experiment. Also, a prototype system to produce fill-in-blank problems and an interactive tutoring system are developed based on the result of the experiment.
ICAI System for Japanese Students to Learn Korean Writing BIBA 845-850
  Kyu-Keon Lee; Katsuhiko Shirai
It is important to find variable cause of composition errors in the composition training for the beginner of Korean language. This paper describes a new intelligent computer assisted instruction (ICAI) system for Japanese beginners to learn Korean composition. Our system supports to clarify the problems caused from the grammatical differences and makes effective learning based on the grammatical similarities. Experimental test was performed to evaluate the practical learning effects and educational effectiveness by this ICAI system, the results of evaluation experiments show the effectiveness of this system.
Designing a System for Cooperative Learning through Social Interaction -- 'AlgoArena': A Learning Tool for Programming -- BIBA 851-856
  Akiko Ide; Hiroshi Kato
AlgoArena is a simulation game of sumo wrestling which provides an effective environment for facilitating the cooperative learning of software programming skills. Its design principles are based on a theory of situated learning which claims that learning is accomplished by participation in a practice community. AlgoArena allows students to form a practice community. The authors have developed a prototype version and implemented it in a pilot experiment with junior high school students. Results indicate the system to be a productive environment in which shared goals and active social interaction among students contributed to the effectiveness of their cooperative learning.
Building Hypermedia for Learning: A Framework Based on the Design of User Interface BIBA 857-862
  S. Mangiaracina; C. Maioli
Research in hypermedia electronic books is a relatively new field that addresses the problem of creating new effective learning environments. However, the lack of well-defined criteria for evaluation and the somewhat marginal effort devoted to the design of multimedia interfaces, has often resulted in the development of a set of multimedia/hypermedia applications which are of dubious value. Studies in HCI and user-centred design can help in the design of hypermedia systems in education. A well-defined set of guidelines, derived from user-interface evaluation principles, can support and guide the design phase and show how to implement the didactic functions effectively. The knowledge domain and tutoring strategy components are then integrated within this framework. The effectiveness of this approach has been successfully tested in the design of our hypermedia system SIDI.

II.9 Learning Environments 2

Interfaces for Learning BIBA 865-870
  J. Sanchez; M. Lumbreras
The current literature emphasizes critical aspects of learning and cognition involved in human-computer interaction. We present a conceptualization for designing interfaces for learning and thinking through the use of modern ideas for building educational software. We address the construction of Hyperstories as a metaphor for enhancing thought and reasoning skills. The advantages of using multimedia for building this type of software, as well as the complexities involved are analyzed and discussed.
Learning Agents: Interface to the Knowledge of a Community BIBA 871-876
  Akira Namatame; Yoshiaki Tsukamoto
This paper presents a model of learning agent and demonstrates how this model will provide a foundation of building the knowledge of a community. The knowledge of a community serves as the repositories for all the accumulated knowledge, and provides a knowledge world for users to explore, to work with, and to discover new knowledge. By learning the structure of knowledge resource distribution patterns as well as users' interests and preferences during system usage, the learning agent customizes knowledge resource retrievals for each user. We also discuss a methodology of building personalized working environments on the knowledge of a community, where, people at physically distributed location, can work together sharing the common knowledge resources.
A Conceptual Framework for the Implementation of MUGICLE System BIBA 877-882
  Hee Jin Cho; Gerardo Ayala; Yoneo Yano
The purpose of this paper is to describe a conceptual framework, called MUGICLE (Collaboration Learning Environment through Multi-Group Interaction), for learning the inflection of a declinable-word and expression of Hangul (Korean language) sentences for Japanese. In this paper we focus primarily on the construction of the system and processing of behavior, the structure of the knowledge base, the knowledge representation, and domain knowledge generation processes. A word class vocabulary for the system has been extracted from standard word sets presented in primary school and the national language textbook of junior high school. This concurrent constructed vocabulary is a set of dictionary of verbs, adjectives, nouns, and so on, all composing 4,000 words. The knowledge base is of a lower rank category of a semantic classification into a word class vocabulary of each, composed of a combination of relations between a semantic network and the semantic categories. Our framework is composed of four components: instructional strategy module, student module, domain knowledge module, and collaborative learning module.
Exploring Self-Developing Models in Interactive, Virtual Learning Environments BIB 883-888
  David A. Smith
An Instructional System for Constructing Algorithms in Recursive Programming BIBA 889-894
  Noriyuki Matsuda; Akihiro Kashihara; Tsukasa Hirashima; Jun'ichi Toyoda
This paper aims at designing an instructional system that teaches recursive programming to a novice programmer. In order to reduce a cognitive load which the novice bears in constructing algorithms in recursive programming, we propose the algorithm and solution for constructing the algorithm. The algorithm represents simple recursion as two iterations. We had an experiment to confirm if the instruction based on the solution is more effective than a programming textbook. As a result, the instruction had a significant effect. We have also designed an instructional system which helps the novice adaptively acquire the solution.
Interacting with a Mediator Agent in Collaborative Learning Environments BIBA 895-900
  Gerardo Ayala; Yoneo Yano
In this paper we present the basic issues in the modelling of a mediator agent in a computer-supported collaborative learning environment. The mediator agent is an intelligent software agent that has been designed based on ideas from agent modelling in distributed artificial intelligence [1] and social learning [2]. Mediator agents in collaborative environments support user interaction based on the communication of the learners' capabilities, goals and commitments. Mediator agents cooperate in the search of learning possibilities for the learners, in order to enhance the relevant collaboration and progress of a networked community of practice.
User Modelling and System Adaptation in the Interactive Anatomy Tutoring System Anatom-Tutor BIBAK 901-906
  Ian H. Beaumont
ANATOM-TUTOR is an anatomy tutoring system for use at university level, combining ITS (intelligent tutoring system) and hypermedia technology. The aim has been to produce a tutoring system which can a) reduce the workload on university staff by automating the presentation of basic knowledge, and b) improve the effectiveness of student study by allowing the system to respond in an individual way to problems encountered by the learner, a feature lacking in the currently widely used drill-and-practice teaching programs. ANATOM-TUTOR is suitable for use both as a reference work and as an active hypertext-based teaching aid, and its self correcting modelling productions enable it to adapt rapidly to the individual user. Its knowledge domain is a section of brain anatomy including the visual system, the pupillary light reflex system and the accommodation reflex system.
Keywords: Adaptive hypermedia, Intelligent tutoring systems, ITS, User modelling, Computer aided instruction, CAI

II.10 Control System for Special Application

Remote Control for Annotated Video BIBA 909-914
  Jeremy Jones; Colin Harris; Ronan Geraghty; Neville Harris
The subject of this paper is the remote control of multimedia applications, using a personal digital assistant (PDA) as a hand held remote control device. Application specific, context dependent user interfaces are transmitted to the PDA for display on its LCD screen. The user interacts with the touch sensitive screen and the input commands are transmitted back to the application. This approach results in a sophisticated and flexible interface which is at the same time easy to use.
Designing a Natural Language Interface for Supervisory Control of Mining Teleoperation BIBA 915-918
  Celestine A. Ntuen; Eui H. Park; Arun A. Setty
In this paper we describe the design of a command language software using expert database. The language structure is contextual and is implemented using a GURU expert system shell. The expert database that drives the operability of the command language utilizes knowledge of mining tasks; and the dialog properties are designed around user intentions and task behaviors.
Expert Inventory Control System for C.I.M: Modelling and its Applications BIBA 919-924
  A. Ajmal
The paper outlines the methodology and knowledge engineering approach to the development of an interactive inventory management system for use in a manufacturing company in the U.K. The prototype system developed with an artificial intelligent (AI) language provides a linkage between management and computer-integrated manufacturing (CIM), and is comprised of four major modules and has a feed back loop which can be used for comparing the actual and estimated values, and automatically updating the inventory records. To test the industrial applicability of the system, data from manufacturing industries have been used. The interactive system which uses a hierarchical decision rules for inventory control domain and uses goal oriented interaction and backtracking interface procedures, has the potential benefits of: better resources planning and management, reduced inventory levels and production lead times, and consistency in maintaining record integrity.
Worm-Type Agents for Intelligent Operation of Large-Scale Man-Machine Systems BIBA 925-930
  Takashi Washio; Masaharu Kitamura
The replacement of conventional analog control and instrumentation systems by digital communication networks is in progress in various industrial large scale systems. The rapid spread of Internet is another trend of public digital communication networks. One of significant features of these communication networks is that each console terminal, network router, sensor, actuator and component in the networks has its own information processing node in form of distributed systems. Each node can perform highly intelligent functions, e.g., self-diagnosis and operation logging, not only its original functions [1]. This architecture has become economically feasible due to the drastic cost reduction of reliable hardware and software of the nodes, and is expected to provide a fundamental structure of future large scale systems.
   The mutual interactions and interferences among parallel operations of the nodes frequently induce complicated situations in the distributed architecture, while they provide efficient and flexible information processing. The conventional centralized and heavily loaded method to control the entire system may not be suitable to manage those situations efficiently. Recently, autonomous software robots (softbot) staying within a node [2] and worm programs able to move among different nodes [3,4] have been proposed for the achievement of efficient and flexible management of the network. These architectures enable detailed network management without load concentration, efficient use of distributed computation resources, and reduction of communication load. However, the generic and efficient architecture to support their high autonomies, adaptability and flexibility has not been established yet, and the studies on the architectures have been limited in some specific applications, e.g., e-mail agents [2,5], diagnostic agents [3,4] and database agents [6].
   The primary objective of this paper is to propose a generic and practical architecture of agents and their network to provide the various types of highly autonomous softbots and worms in terms of man-machine systems. It supports rational operations of agents and their appropriate self-organization under resource bounded conditions. The second objective is to demonstrate the efficient performance of this architecture in some practical applications of control, diagnosis and message exchange providing a new environment of large scale systems for human operators.

II.11 Cognitive Model in Process Systems

Development of Operator Cognitive Model in Analysis Support System for Man-Machine System Design Information BIBA 933-938
  Takashi Nakagawa; Kazunori Sasaki; Toru Itoh; Hidekazu Yoshikawa; Makoto Takahashi; Kazuhiro Kiyokawa; Akira Hasegawa
An integrated software system has been under development with aims at analyzing and evaluating the effectiveness of man-machine system design, by computer simulations from various viewpoints of human-factors. In this paper, the configuration of a distributed simulation system is first introduced, followed by an explanation of how the operator simulator model is organized by a Petri net model. Also presented is an example simulation of a plant anomaly diagnoses procedure using the Petri net model in its current state of development in the initial phase of software development.
Computer Simulation System of Cognitive Man-Machine Interaction in Accidental Situation of Nuclear Power Plant BIBA 939-944
  K. Yoshida; M. Yokobayashi; K. Kawase; F. Tanabe
A prototype version of computer simulation system has been developed at JAERI (Japan Atomic Energy Research Institute) to provide analysts with detailed information on cognitive process of operator in accidental situation of NPP (Nuclear Power Plant). The simulation system consists of a plant model and an operator model. A model of operator cognitive behavior is based on the decision ladder model of Rasmussen, and is implemented using AI-technique of a distributed cooperative inference method with a so-called blackboard architecture. Rule-based behavior is simulated using If-Then type of rules. Knowledge-based behavior is simulated using knowledge representation with Multilevel Flow Model (MFM) and qualitative reasoning method. Some trial simulations of incident have been performed to verify the modeling of operator cognitive behavior in NPP accident.
Using Simulated Human Model for Evaluating Human-Machine Interaction BIBA 945-950
  Yushi Fujita; Ichiro Yanagisawa; Hiroshi Sakuda; Kazuhiro Kiyokawa
Human-machine system can only be understood correctly when it is treated as a joint system. Neither humans nor machine can determine its performance in isolation. It is the interplay between humans and machines that determines the performance. Hence, it is a prerequisite to look into context-driven interactions between humans and machines when attempting to evaluate human-machine systems. Because of the dynamic nature of the interactions, experimental approaches have widely been used for the analysis of the interactions. However, they tend to be associated with a number of practical constraints that sometimes make them difficult to apply. As a supplemental means, simulated human models can be utilized as a useful tool for analyzing the interactions. Cognitive and Action Modeling of Erring Operator (CAMEO) is such a human model which features its ability to simulate some typical erroneous tendencies associated with cognitively demanding tasks such as nuclear power plant operation. This paper describes the technical foundation of CAMEO.
Experimental Study on the Operators' Cognitive Activities Based on the Diverse Information BIBA 951-956
  Makoto Takahashi; Osamu Kubo; Akira Yasuta; Hidekazu Yoshikawa; Kazunori Sasaki; Toru Itoh; Masayuki Matsumiya; Takeharu Sakaue; Kazuhiro Kiyokawa; Akira Hasegawa
Experimental study on the operators' cognitive activities had been performed using nuclear power plant simulator. The specific feature of the present study is that the state of the operator during the interaction through the man-machine interface is analyzed based on the diverse information; 1) Estimated cognitive state, 2) Estimated mental work load, 3) Operational sequence history, 4) Questionnaire and 5) Video record. These informations are integrated and graphically represented on the form of Chronological Diagram, by which the operators' dynamic cognitive behavior can be analyzed. Results of the simulator experiments, in which three expert operator participated, demonstrated that the present method can provide the valuable information concerning the operators' cognitive behavior during the task of anomaly diagnosis.
Analysis of Operator's Diagnostic Behavior using Computer Simulation BIBA 957-962
  K. Furuta; M. Takahashi; H. Yoshikawa; K. Sasaki; T. Itoh; M. Matsumiya; T. Sakaue; K. Kiyokawa; A. Hasegawa
Computer simulation was performed on a cognitive experiment of abnormal event identification in a nuclear power plant. The process of identification has been modeled as similarity matching of symptomatic patterns and hypothesis validation by symptom observation. As a result of comparison, the simulation could explain behaviors of the subjects generally well, and some characteristics of operator's cognitive processes have been revealed.

II.12 Interface Design and Evaluation in Process System

An Object Oriented Methodology for Man-Machine Systems Analysis and Design BIBA 965-970
  A. Mahfoudhi; M. Abed; J-C. Angue
Despite the recent progress in the domain of Man-Machine Interface engineering, several problems concerning the incompatibility between the information presentation to the user and his cognitive representation are still present. This paper presents a new Task Object Oriented Description methodology (TOOD), specially adapted to the taking into account of the human factors for the specification of the Man-Machine Interfaces (MMI). A concrete application of this methodology was presented in the air traffic control context.
Cognitive Functions and Automation: Principles of Human-Centred Automation BIBA 971-976
  Erik Hollnagel
This paper describes the principles of human-centred automation from the point of view of cognitive systems engineering. Five main aspects of automation are identified: level of automation, interface, computer support, training, and task allocation. A matrix is proposed which describes the dependencies between these five aspects, thus maintaining the joint systems perspective.
Evaluation of Ecological Interface Design BIBA 977-982
  Osami Watanabe; Katsuhisa Takaura; Yushi Fujita; Yoshio Hayashi
It has increasingly been recognized that operators working in a highly automated large-scale human-machine system may be exposed to cognitively demanding situations in unanticipated emergency. A new display design called ecological interface design (EID) is proposed by several researchers, which they claim is supportive especially in such difficult situations. The EID features model-based parameter selection and graphic presentation, which fit the cognitive characteristics of operators. This paper presents the results of a comparative evaluation of an EID display.
Task Modelling Using Object Petri Nets BIBA 983-988
  S. A. Kaddouri; H. Ezzedine; J. C. Angue
This study aims at developing a methodology for interfaces specification which is most concerned with different working situations that the operator could deal with. For each situation, a task model is constructed. We distinguish two parts: one draws up the goals that the operator tends to reach as a response to a specific situation, the other, draws up the different actions that the operator has to trigger in order to satisfy the former requirement. The description of these two parts is done using object oriented Petri nets (OPN).
Simulation Study of an Ecological Interface for Nuclear Power Plants BIBA 989-994
  A. Sakuma; J. Itoh; E. Yoshikawa; K. Monta
Application of the ecological interface design to nuclear power plants is presented and the method for its evaluation is discussed. Starting from a concept for intelligent man-machine systems, the ecological interface is introduced as the concept's major ingredient and application to the plant shutdown is described. It is desirable for the evaluation of complex man-machine systems to be concurrent with the design of such systems. A hybrid evaluation, i.e., a combination of an analytic and an empirical evaluation seems promising. The taxonomy for work analysis by Rasmussen provides a good framework for this evaluation.
The "Bird's Foot" Integrated Graphical Interface for NPP Operation BIBA 995-998
  N. Moray; B. G. Jones; P. M. Sanderson; D. V. Reising; S. Shaheen; J. Rasmussen
The traditional single-sensor-single indicator display is poorly matched to the cognitive abilities of operators, especially for large and complex systems. Our research aims to provide direct perception displays which will greatly reduce the cognitive load on the operator and allow the use of perceptual rather than cognitive mechanisms to support start-up, state diagnosis and fault management. In particular we describe a system of displays which supports movement by the operator up the abstraction hierarchy as the plant is brought on line. The aim is to minimise the cognitive load on the operator by providing information only at the level at which the operator is currently thinking, while at the same time showing how close the system is to operational limits.

II.13 Human Factors in Nuclear Power Plants

Development of a Research Simulator for Human Factors (1) -- An Outline of Development of the Research Simulator -- BIB 1001-1006
  Ryutaro Kawano; Shinya Shibuya; Katsumi Nagata; Masahiro Yamamoto
Development of a Research Simulator for Human Factors (2) -- Hardware and Software of the Research Simulator -- BIB 1007-1012
  Shinya Shibuya; Ryutaro Kawano; Katsumi Nagata; Masahiro Yamamoto
An Intelligent Computer Aided Instruction System for Plant Operators to Study Emergency Procedures BIBA 1013-1018
  Hiroshi Ujita; Takeshi Yokota; Naoshi Tanikawa
Intelligent CAI system has been developed which makes it possible to provide consistent education in plant operation and also plant behavior for professionals like shift supervisors. Satisfaction of intrinsic motivation was tried by representing instruction according to the learners' level. The student model, which is common to teaching course for presenting text knowledge of emergency procedures and the indicating course for actual plant behavior and procedures using the plant simulator, was derived from a hierarchical function model which is a goal oriented mental model of a plant operator. The understanding level of each node (element of a function) in the model is evaluated by personal history conditions calculated from both the tutoring record of the node and the understanding level of the connecting nodes.
An Interactive Support System for Incident Analysis in Nuclear Power Plants BIBA 1019-1024
  Wan C. Yoon; Yong H. Lee; Young S. Kim
HPES, a managerial error analysis system to investigate the types and causes of human errors in nuclear power plants (NPP), has been widely adopted in U.S. and many other countries. To make the reporting more effective, the chronological description of the incident has to include the cognitive activities of the operators, and the analysis of human errors needs to be explicitly related to the human cognitive process. This paper presents a computerized support system for the analysis of human errors in NPP based on a cognitive model. The most intelligent part of the support system interactively aids the analyst in composition of event sequences by checking possible omissions of important facts and logical deficiencies against a decision making model. It also helps the analyst to relate error classification and factor analysis parts directly to the event description. As a result, the analyst not only can conduct the analysis more easily but also can develop a report of a better quality that reveals cognitive causations among human activities and system states.
Development of an Operator Training Support System for Nuclear Power Plants BIBA 1025-1030
  Tsutomu Ohtsuka; Katsuji Yaguchi; Sadanori Yoshimura; Yoshinori Takada; Kunio Noji
An operator training support system (OTSS) for nuclear power plants is developed based on the analysis of operator performance under simulated plant abnormal and accidental conditions at the BWR Operator Training Center.
   This system automatically evaluates the operation performed by trainees and provides information on the problems in trainees' operation in real-time by using evaluation logic stored in a knowledge base.
   The operator training support system improves the accuracy and efficiency of evaluation in training.

II.14 Operation Aids Plant Diagnosis

Interactive Plant Management with Real Time Conformity Checking: The ImagIn Project BIBA 1033-1038
  David Haubensack; Pierre Malvache; Philippe Valleix
This paper describes a research project of the French CEA, addressing to the evolutions in plant operation apt to bring perceptible and assessable improvement in the operational safety [1].
   Many mistakes in plant operations are due to a discrepancy between the "mental representation" of the plant by the operators and the actual plant state: this is often due to lack of information provided to operators, particularly on the modifications of the plant, either temporary or definitive. This can also originate in an inconsistency between the operational procedures and the actual state of the plant, due to these modifications. The maintenance of a coherent and unique representation of the plant for all the actors (human or computerized) of plant operations is the objective of the ImagIn project [2].
Development of an Attractive Computer Assisted Instruction System for Plant Operators BIBA 1039-1044
  Keiko Mutoh; Ryutaro Kawano; Hiroshi Ujita; Takeshi Yokota; Ryuji Kubota
A new concept CAI system for nuclear power plant operators has been developed. The main focus of the development was on operator's motivation. The system was accepted considering attractiveness with some requirements of operators extracted to improve the system.
   The need for voluntary learning is growing ever greater for operators to maintain their skills. We will add some functions to improve the CAI system to help elicit the operator's voluntary learning.
The Design of Computerized Procedure Presentation for Nuclear Power Plants BIBA 1045-1049
  Yuji Niwa; Erik Hollnagel
New methods of information presentation and interface design are changing the conditions for work in the modern nuclear power plant (NPP) control room. One area receiving considerable attention is that of Emergency Operating Procedures (EOP), which play an essential role in NPPs. This paper presents the results of a project which made consistent use of cognitive engineering and cognitive ergonomics principles to specify a prototype system for computerized procedure presentation in a NPP. The system is scheduled for completion and field evaluation in the summer of 1995.
Development of the On-Line Operator Aid SYStem (OASYS) using Rule Based Expert System and Fuzzy Logic for Nuclear Power Plants BIBA 1051-1056
  S. H. Chang; H. G. Kim; S. S. Choi
In this paper, the On-line Operator Aid SYStem (OASYS) has been developed to support operator's decision making process and to ensure the safety of a nuclear power plant by timely providing operators with proper guidelines according to plant states. The OASYS consists of four systems such as signal validation and management system (SVMS), plant monitoring system (PMS), alarm filtering and diagnostic system (AFDS) and dynamic emergency procedure tracking system (DEPTS), and is based on a rule based expert system and fuzzy logic. The rule based expert system is used to classify the pre-defined events and track the EOPs through data processing and the fuzzy logic is used to perform the prognostic diagnosis and to evaluate the qualitative fuzzy criteria used in the EOPs. Evaluation results show that the OASYS is capable of diagnosing plant abnormal conditions and providing operators appropriate guidelines with fast response time and consistency. Currently the OASYS is installed in the real time full scope simulator for validation.
Development of Abnormal Signal Separation Method in Nuclear Power Plant BIBA 1057-1062
  Yoshikazu Umeda; Hiroshi Inujima
In an nuclear power plant various data are periodically collected from the installed sensors. These data are used as criteria for the safety confirmation of the plant and the replacement of machines. However, in actual practice these data are operated by the standard based on the threshold values given in the "Operation Manual," etc. On the other hand, when noise gets mixed up with these data, the control value for plant operation apparently gets increased, and is likely to cause misjudgement. Moreover, the increase in control value for operation may be giving some information regarding the plant. It has long been considered that it is possible to obtain some new information about the plant; the information not included in the "Operation Manual" by putting these collected and stored data to appropriate analysis. This information refers to the deterioration, change in condition, etc. of the machine, and is expected to contribute to higher maintenance support. However, expert as the data controller of an atomic power plant may be regarding the "Operation Manual," he is not a specialist of data analysis, so that the enormous amount of data are left almost unanalyzed. Under such background, research has been made to upgrade the maintenance support level of the atomic power plant. This report describes the concept of "maintenance support high-grade expert system," on the basis of the data analysis made by using a computer, and deals with the results obtained through application of the examples after trial-manufacture of the prototype.

II.15 Intelligent and Adaptive Systems

Visual Object Recognition System with Selective Attention Mechanism BIBA 1065-1070
  Keiko Ishihara; Shigekazu Ishihara; Mitsuo Nagamachi
We can identify just interested objects quickly even in the scene that has many objects. Selective attention is the powerful function of human visual processing for adapting changing environment. The 3-D object recognition system presented in this article has attentional mechanism that is suggested from psychological findings. Preattentive peripheral vision receives a gray-scale image low resolution, then it marks small regions to attend that have significantly larger luminance than neighbor. Attentional foveal process detects edges in the small attended area and then specifies features for retrieving object models. The object models are described by geons (Biederman, 1990) for fast flexible recognition.
User Modelling for Adaptable Interface Design BIBA 1071-1076
  D. Akoumianakis; C. Stephanidis
A prototype user modelling module (UMM) has been designed and implemented which supports the automatic derivation of lexical level user interface adaptability rules. UMM is part of a new high level user interface development environment, currently under development, which facilitates the construction of unified interfaces for interactive applications, accessible by different user groups, including people with disabilities. UMM supports the acquisition and storage of information about target users and produces a set of rules for adapting the user interface at the lexical level; this comes as a result of a three-phase process, namely reasoning about, selection of, and decision on the optimal interaction techniques and devices.
Adaptive Human Computer Interfaces for Supervision Systems BIBA 1077-1082
  E. Furtado; Ch. Santoni; Ph. Francois
In this paper the authors show both the adaptation and the generation processes of supervision system interfaces. For the adaptation process, they take into account the operator, the task and the interaction models. For the generation process, they show how to generate interfaces automatically from the task model into a formal multi-agent representation.
Hybrid Machine Learning: Myth and Reality BIBA 1083-1088
  Vassilis S. Moustakis; Gavriel Salvendy
An important issue to consider when applying Machine Learning (ML) in a real world task is the selection of a system, algorithm or approach which should be used. In this context coupling of the right ML approach with the task at hand is not trivial. This paper reports the preliminary results of a research which targeted to coupling ML approaches with generic intelligent tasks. Preliminary analysis makes it clear that in most of tasks application of a single ML approach is not satisfactory and that hybrid formations are necessary.
An Interactive Modeling Technique for Hair Styles BIBA 1089-1094
  Yoshiaki Kawase; Takeshi Maeda; Yoshio Ohno
The display of human image is important for improving the user interface. It is still difficult, however, to generate realistic human images even for the advanced computer graphics of today. One of the reasons is the difficulty in shaping the specification and rendering of human hair.
   The specification of the shape of the object to be handled in the computer is called (geometric) modeling. The modeling of hair has relied mainly on handwork, and is very time-consuming. Here, we propose an interactive modeling technique using a 3D digitizer and an interpolation method for the generation of large amounts of hair. These methods drastically reduce the work for hair style modeling.

II.16 Machine Learning

Managing Machine Learning Application Development BIBA 1097-1102
  Vassilis S. Moustakis
Machine Learning (ML) represents an advanced technology and its effective implementation in the workplace is not a trivial task. This paper overviews areas of applications for ML and discusses management issues that are relevant to ML application development.
Automated User Modeling for Intelligent Interface BIBA 1103-1108
  Kenichi Yoshida; Hiroshi Motoda
The analysis of the user behavior is one important function of the intelligent, user interface because, by analyzing the user behavior, it becomes possible to understand the user intention and release the user from tedious tasks which are often required to use a fast hut low-level interface. The acquisition of the user behavior model is crucial. Most studies meant to realize an intelligent interface system only analyze superficial user behaviors, from which to automate the repetitions. Their user models tend to be simple and do not reproduce the behavior well enough. This paper presents a new framework that analyzes the computational processes activated by the user commands to build the user behavior model. An important, feature of the proposed framework is the analysis of data dependency between the user commands. A user adaptive interface system, Clipboard, was developed to show the adequacy of this framework. It analyzes the I/O relationship between applications in the past, task history, selects the next application, and creates scripts which enable complex task execution by a single command.
Predicting Pilot Bid Behavior with Genetic Algorithms BIBA 1109-1113
  Pieter Adriaans
We describe our experiences in building CAPTAINS (Crew Availability Planning and Training System) for the Royal Dutch airline company KLM. CAPTAINS is a complex application that enables a planner to maintain strategic, tactic and operational models of pilot populations. A major problem in building the short-term planning algorithms for CAPTAINS was prediction of pilot bid behavior. Twice a year pilots may express their preference for new seats (functions) and KLM is obliged to give a new seat to the most senior officer who is qualified. However in some cases the bids are not yet known for the season for which one is planning. If a pilot changes his bid, this can influence the planning substantially, so that a correct prediction of pilot bid behavior has vital importance for a good plan. Using genetic algorithms, we were able to produce rules that predict pilot bid behavior with an acceptable level of accuracy.
   Experience with CAPTAINS shows that information technology can realize a substantial reduction in overall manpower costs and improve the service level of the planning department. To realize planning applications of this kind, a number of heterogeneous techniques are required, including operations research, relational database management, rule-based reasoning, machine learning, graphical user interfaces and constraint handling. The development of applications of this complexity demands new project management techniques and new design tools. Here we will focus on the machine learning aspects of the CAPTAINS project.
Knowledge Acquisitions from Large Databases using Machine Learning Techniques BIBA 1115-1120
  Yasubumi Sakakibara
The rapid growth of data in large databases such as text database, scientific database requires efficient computer methods for automating analyses of the data with the goal of acquiring knowledges or making discoveries. Since the analyses of data are generally so expensive, most parts in databases remains as raw, unanalyzed, primary data. Technology from machine learning theory will offer efficient tools for the intelligent analysis using "generalization" ability. Generalization is an important ability specific to inductive learning which will predict unseen data with high accuracy based on learned concepts from training examples. We will demonstrate the effectiveness of our approach where generalization ability is applied to predicting and analyzing primary data and extracting knowledges from database by presenting some our results on text database analysis and biological sequence analysis.
Machine Learning Methods for Intelligent Assistant Systems BIBA 1121-1126
  Jurgen Herrmann
Intelligent assistant systems provide an adequate organization of human computer interaction for complex problem solving. Machine learning methods can provide significant support for assistant systems. The learning apprentice system COSIMA is presented that acquires knowledge about single problem solving steps.
Machine Learning in Human-Robot Interaction BIBA 1127-1132
  Kazuo Hiraki
This paper discusses a possible learning architecture for human-robot interaction. Intelligent robots need to share knowledge with human beings for flexible interaction. However, the gap between low-level sensory data and abstract human knowledge makes it difficult to pre-encode the robot behavior against the human's various complex demands. To avoid this problem, robots should have a learning mechanism which is applicable to sensory/perceptual information. In order to overcome the gap between the low-level sensory data and higher-level concept description, we developed a learning system, called ACORN-II. ACORN-II uses a method called feature abstraction, which dynamically defines abstract sensors from primitive sensory devices and makes it possible to learn appropriate sensory-motor constraints. The paper demonstrates ACORN-II with some experimental results and discusses the future direction of human-robot interaction.

II.17 Intelligent Interface Design

User Interface Design using Mental Model Transfer BIBA 1135-1140
  Yasufumi Kume; Hyun Seok Jung; Nozomi Sato; Eiichi Bamba
In this paper, the concept of mental model transfer is introduced to overcome the problem of accurate transfer from system developers to end users and that of reducing the software developing time. In addition, mental model will be classified into "model in mental world" and "model of mental world". The former model is introduced to design user interface of a shell for diagnostic system. This paper describes how to transfer designer's mental to maintenance engineer. The effect of this method is discussed.
Usability Testing for a Tele-Radiology Workstation BIBA 1141-1146
  Nahm Sik Lee; Jae Hee Park
An empirical study was conducted to provide design guidelines for the user interface (UI) of a computer-based tele-radiology workstation, MIDAS (Medical Image Display and Archiving System). Under eight task-scenarios, subject performances such as completion time, error and user interactions were recorded by auto-logging method. After the each test session, subjective satisfactions were also asked using a questionnaire. Considering the collected data, current design defects on screen layouts, wordings, operational procedures were analyzed. Also, distinctive types of errors were classified and the causation of error was analyzed with regard to the UI attributes. The performance metrics showed that the doctors (novice in computers) require twice as much time as the college students (expert in computers) to acquainted with the new workstation. Finally, UI design guidelines for a tele-radiology workstation were suggested for the better design of a future system.
Adaptive Intelligent On-Line Help System BIBA 1147-1152
  Chau-Hung Wang; Chung-I Lee
This paper introduces the architecture of an Adaptive Intelligent On line Help System (AIOHS). Five constructive components are needed to create such system: (1) interface manager; (2) knowledge base; (3) inference engine; (4) knowledge checker; and (5) help information base. An AIOHS to be applied to some portion of text editor is constructed in this research. This proposed help system is designed to provide help to different classes of users. One strength of the system is the knowledge base and inference engine which can detect errors and identify the object user want to get. Another strength is the knowledge checker which generates different help strategies determined by the level of the user's operating performance. An experiment is conducted to verify the AIOHS, the results indicates that AIOHS has good efficiency in problem solving.
User-Friendly Interface Development for CAD-Based Engineering Data Management Systems BIBA 1153-1158
  Ting-Kou Peng; Amy J. C. Trappey
A system designed to help engineers define and manage product engineering data is called an Engineering Data Management system (EDMS). Although a lot of CAD tools for 3-D design applications are capable of modeling objects using personal computers, there are still obstacles to manage huge existing product engineering data or define a new product comprehensively and efficiently. Further, these CAD systems often have functional limitation for specialized applications. This paper will develop an EDMS to manage existing products data and define new products. We focus on the development of a user-friendly interface for EDMS in order to achieve the objective of a user-oriented design and engineering environment.
Interfaces for Synthetic Speech Creation BIBA 1159-1164
  Hiroshi Hamada; Ikutoshi Mizumoto; Jin'ichi Chiba
Text-to-speech synthesis is widely used in the interfaces of phone-based services such as telephone ordering systems and computer interfaces such as read-out and collate systems for text edited by word-processors [1]. However, since synthetic speech is generated automatically by fixed acoustic and linguistic processes with average values, it is monotonous and not sufficiently intelligible. Although some speech synthesis systems can control the speaking speed and fundamental frequency, it is difficult for users who are unfamiliar with speech technology to synthesize the desired speech quality. In some applications, information providers using text-to-speech synthesis want to change the speech style to match their intention. For example, they may want to emphasize specific words that are important to the listener. This indicates the necessity of developing a simple user-interface that makes it unnecessary for the user to directly control the physical speech parameters; the interface mediates between the user and the control of the parameters.
   The purpose of our research is to develop simple interfaces that allow the user to create synthetic speech using terms clear to the user such as "strong emphasis" and "more slowly". This paper introduces an enhanced graphical user interface (GUI) for altering the prosodic features of synthetic speech. We also describe the design procedure used and the design results of simple interfaces for modifying speech style in terms of users intensive direction.
Sensitive Screen HCIs in Recent Siemens Telecommunication Products BIBA 1165-1170
  R. Schoeffel
This paper presents recent HCI research at Siemens for an intelligent Home Assistant. It presents an example for product development emphasizing man machine communication with more intelligence. Goals and strategies of Siemens research towards intelligent human computer interfaces are outlined.