Proceedings of the workshop "Intelligent Educational Systems on the World Wide Web",
8th World Conference of the AIED Society, Kobe, Japan, 18-22 August 1997

From adaptable to adaptive interface for distance education

Aude Dufresne
Dept. of Communication, University of Montreal
dufresne@iro.umontreal.ca

1 Usable interfaces for distance education

We are developing in the context of the Canadian TeleLearning Network of Center of Excellence a distance education environment on the net. The system will offer not only hypermedia content on the net, but different tools for conferencing, planning meeting, white board, videoconferencing, scenario and knowledge representation, annotations (Fx - Nomino), etc. The system is very complex, so we have to develop supportive interfaces that will adapt to the user, unfolding gradually as he gain in expertise. A support system to advise the learner on his use of the system is being developed, which for now use the help facilities of the Windows'95 environment, but which will eventually drive the interface and give the user graphical cues (Lee & Lehman, 1993) using text, graphics and animations, voice or eventually force feedback and guiding. The architecture of the system is partly local and partly shared, it is an integration of commercial Office'97 softwares and applications running in Visual Basic, C++, HTML and Java applets.

Figure 1 show the different tools accessible to the user, Some tools are Office'97, others are programmed in Visual Basic. A central information system works locally to facilitate the communication among local applications and also for communication to the server of the user activity. The web serves for the communication and diffusion of the user's model and activity in the environment. Using Java applets or VB, feedback is given to him of his (progression in the notes he takes on the different elements of knowledge, on the web or in Word), and others activity (presence of others online, attitude toward workgroups). The user's model keeps the navigation path and his preferences; it is located both locally and on the server.

2 Collaborative adaptation

In fact, the context of distance learning, is one where the system is not at the center of the interaction, the teacher, the other learners are also there so the system is seen more as a tool, then as the only interaction. As such, it cannot assume the full direction of the activity, it must be transparent to the user goals and activity and cannot be too impredictible. In this context what the user needs, is not as much a feedback on his progression in the content but a guide as to how he may use them to achieve the propose objectives of the learning. Interface adaptation can be disruptive and also there must be some synchrony between the students' progression which suggest some goals and structures in the interaction that are external to the student model. In this context a too intrusive system can be disruptive to what ever objectives the users may have on his own. The concept of collaborative adaptation is a very important area of research (Kay, 1995), where the user defines how much and what can be adapted (Kaplan & al 1993). In our system the user specifications are set directly, as he may change the system contextual help, fine tuning the level and types of system's interventions: delay before the identification of objects, suggestion of tasks, etc.

Another aspect of the user directing the support is to offer him the opportunity to specify is goals. The learner may choose among a set of general intentions, and the system then suggest him specific way to realize them in the different tools of the Virtual Campus: Explore, Plan, Search, Evaluate his own progression. He then follows him depending on what environment he choose and give specific advice depending on where he his as to on what he should focus his attention and activities.

Figure 1 (full-size)

3 Adaptable contextual support

Help messages are designed using Windows'95 help facilities and Robot Help, and an Access database for representing rules of intervention. A model of support depending on the context and on the user's trace is being developed. The system offers contextual help depending on the user historic navigation in the environment. number of visits, intentions identified by the users, path in related environment are used to define which should be given to the user in his use of the environment. This model is built upon the link between activity theory and interface design, which state that the user must first be told Goals (Why ?), elements of the artefact (What ?) and methods on How ? he may use the environment. Advice is given to the user, as suggested by Browne (1990) , on what object he should focus in the interface and his activities, depending on his progression. Theories on the development of expertise, on perception and on motivation are also integrated in the definition of the type and degree of interventions of the system: not at the beginning, never twice, when it really makes a difference.

The user can control these interventions of the system, simply turning the system off or using preferences sliders to modify different types of help. We intend to evaluate using observation, trace analysis and questionnaires, the usability of different modalities of help in the environment and to compare how the user turn them off or on at the beginning of the session vs later as he gain expertise. Such a mixed adaptive environment is the best way for the environments to be really adapted to the user learning situation (Brusilovsky 1996), but also to his evolving wish. From adaptable to adaptive functions Adapting the interface might be a tedious task, which learners might not have time to do, so we intend to integrate adaptive functions that will integrate both global and individual users' feedback on the help system, to adjust future help both in general and individually. Such learning mechanisms we think are important to guarantee adaptation of systems to longer learning situations. These adaptive support functions which are now only to support the usability of the system, will be matched to advisors aiming at pedagogical or collaborative support. The adaptive functions will make it possible to adapt the system to each user in these three area separately. In fact, the context of distance learning, is one where the system is not at the center of the interaction, the teacher, the other learners are also there so the system is seen more as a tool, then as the only interaction. As such, it cannot assume the full direction of the activity, it must be transparent to the user goals and activity and cannot be too impredictible. In this context what the user needs, is not as much a feedback on his progression through the content (Dufresne, 1991; De La Passardiere & Dufresne 1992) but a guide as to how he may use them to achieve the propose objectives of the learning (Dufresne,1997).

Figure 2

Interface adaptation can be disruptive and also there must be some synchrony between the students' progression which suggest some goals and structures in the interaction that are external to the student model. In this context a too intrusive system can be disruptive to what ever objectives the users may have on his own. The concept of collaborative adaption is a very important area of research (Kay, 1995), where the user defines how much and what can be adapted (Kaplan & al 1993). In our system the user specifications are set directly, as he may change the system contextual help, fine tuning the level and types system's interventions: delay before the idenfication of objects, suggestion of tasks, etc. Another aspect of the user directing the support is to offer him the opportunity to specify is goals. The learner may choose among a set of general intentions, and the system then suggest him specific way to realize them in the different tools of the Virtual Campus: Explore, Plan, Search, Evaluate his own progression.

Figure 3

He then follows him depending on what environment he choose and give specific advices on what he should focus his attention and activities.

Adaptable contextual support

Help messages are designed using Windows'95 help facilities and Robot Help, and an Access database for representing rules of intervention. A model of support depending on the context and on the user's trace is being developed. The system offers contextual help depending on the user historic navigation in the environment. number of visits, intentions identified by the users, path in related environment are used to define which should be given to the user in his use of the environment. This model is built upon the link between activity theory and interface design, which state that the user must first be told Goals (Why ?), elements of the artefact (What ?) and methods on How ? he may use the environment.

Advice is given to the user, as suggested by Browne (1990) , on what object he should focus in the interface and his activities, depending on his progression. Theories on the development of expertise, on perception and on motivation are also integrated in the definition of the type and degree of interventions of the system: not at the beginning, never twice, when it really makes a difference. The user can control these interventions of the system, simply turning the system off or using preferences sliders to modify different types of help.

We intend to evaluate using observation, trace analysis and questionnaires, the usability of different modalities of help in the environment and to compare how the user turn them off or on at the beginning of the session vs later as he gain expertise. Such a mixed adaptive environment is the best way for the environments to be really adapted to the user learning situation (Brusilovsky 1996), but also to his evolving wish. From adaptable to adaptive functions Adapting the interface might be a tedious task, which learners might not have time to do, so we intend to integrate adaptive functions that will integrate both global and individual users' feedback on the help system, to adjust future help both in general and individually. Such learning mechanisms we think are important to guarantee adaptation of systems to longer learning situations. These adaptive support functions which are now only to support the usability of the system, will be matched to advisors aiming at pedagogical or collaborative support. The adaptive functions will make it possible to adapt the system to each user in these three area separately.

4 From adaptable to adaptive functions

Adapting the interface might be a tedious task, which learners might not have time to do, so we intend to integrate adaptive functions that will integrate both global and individual users' feedback on the help system, to adjust future help both in general and individually. Such learning mechanisms we think are important to guarantee adaptation of systems to longer learning situations. These adaptive support functions which are now only to support the usability of the system, will be matched to advisors aiming at pedagogical or collaborative support. The adaptive functions will make it possible to adapt the system to each user in these three area separately.

References

Arcand, J.-F. (1995). An Artificial Neural Network for the Ergonomic Evaluation of a Human-Computer Interface. , Intern. conf on Tools with Artificial Intelligence, IEEE, New Orleans, Louisiane(pp. 606-608).
Beaumont, I., & Brusilovsky, P. (1995). Adaptive Educational Hypermedia: From Ideas to Real Systems. , ED-Media'95, Graz Austria.
Browne, D., & Totterdell, P. (1990). Adaptive User Interfaces. London: Academic Press, 227p.
Brusilovsky, P. (1996). Methods and Techniques of Adaptive Hypermedia. user Modeling and User-Adapted Interaction, 6, 87-129. De La Passardière, B., & Dufresne, A. (1992). Adaptative Navigational Tools for Educational Hypermedia. In I. Tomek (Ed.), Computer Assisted Learning (pp. 555-567). Berlin, New York: Springer-Verlag.
Dufresne, A. (1997). HYPEREXP: A Generic Hypermermedia and Knowledge Based interface to Teach Decision Making. Ed-Media 97.
Dufresne, A., & Tremblay, I. (1991). Modelling Distributed Knowledge to Support Learning Environments in Physiology and Computer Science. International Conference on Simulation in Engineering Education. Simulation Series, 24(2), 185-189.
Elkerton, J., & Williges, R. C. (1989). Dialogue Design for Intelligent Interfaces. In P. A. Hancock, & M. H. Chignell (Ed.), Intelligent Interfaces: Theory, Research and Design (pp. 213-264). Amsterdam: North Holland.
Kaplan, C., Fenwick, J., & J., C. (1993). Adaptive Hypertext Navigation Based on User Goals and Context. User Modeling and User-Adapted Interaction., 3(3), 193-220.
Kay, J. (1995). The UM Toolkit for Cooperative user Models. User Modeling and User-Adaptated Interaction, 4(3), 149-196.
Lee, Y. B., & Lehman, J. D. (1993). Instructional Cuing in Hypermedia: A Study with Active and Passive learners. Journal of Educational Multimedia and Hypermedia, 2(1), 25-37.
Linard, M., & Zeiliger, R. (1995). Designing Navigational Support for Educational Software. , EWHCI'95, Springer, Moscow, Russia(pp. 63-78).