音声ブラウザご使用の方向け: SKIP NAVI GOTO NAVI

Web Posted on: August 4, 1998


Developing an Adaptive Learning Environment for the Disabled

Jaakko Kurhila
Erkki Sutinen
Sampo Jokinen
Ran Nyman Pasi V„is„nen

Department of Computer Science, P.O. Box 26, FIN-00014 University of Helsinki
tel: +358 9 708 44664, fax: +358 9 708 44441
email: kurhila@cs.helsinki.fi

 

1. Summary

We have designed a framework for a computerized adaptive multi-person learning environment to be used in special needs education. The framework is a result of a software development scheme where solutions have been designed and evaluated in collaboration with special teachers, neuropsychologists, computer scientists and future users.



| Top |

2. Introduction

Computer-aided learning has been under an intense research for decades. Special needs education provides a particular challenge to computer-aided learning, since every learner is a unique one. However, state-of-the-art achievements in computer science are not harnessed to the use of special needs education.

When targeted to special needs, educational software is usually for visually impaired persons. Much attention has been paid to the design of alternative user interfaces, following the tradition of assistive technology (see e.g. Ramstein et al. 1996). On the other hand, the inside of educational software, i.e., the pedagogically sound content and its technologically advanced implementation, is too often forgotten. However, utilizing the full potential of computer reaches new target groups for computer-aided learning. We have focused on applying computers to support the learning of those children having difficulties in mental programming (e.g. Welsh et al. 1991). When simplified, mental programming refers to one's ability to compose problem solving strategy and uphold the attention.

We introduce a multiprofessional development scheme of educational software. The scheme was applied in the design and implementation of an intelligent and adaptive computer-aided learning environment to be used in special needs education. The aim of our project is to design a networked learning environment in which various topics can be taught and learned. The learning environment provides the learner with exercises which adapt to his/her capabilities. In addition to learners, the same environment can be used by for example teachers, occupational therapists, and neuropsychologists. Also, the communication between the learners is possible to enable cooperative learning. Eriksson et al. (1997) give a more detailed description of the environment.



| Top |

3. Adaptive prototype

We have implemented the first prototype of the learning environment described above. Currently, it is for one learner, and it is used in teaching basic arithmetics. The level of the content and the functionality as well as the visual design are refined in collaboration with special teachers and neuropsychologists. The design is based on a neuropsychological model of the functionality of the human brain and the capability of learning (Korkman 1988).

The prototype can adapt to the learning process according to the learner's needs so that each user has his/her own individualized learning path. The prototype adapts to the user's capabilities by raising or lowering the difficulty level of the exercises. Before lowering the difficulty level, the program tries to help the user to solve the exercise. By guiding the learner through the exercise space, every learner is bound to have his/her own learning path.

The way of helping the learner is totally decidable by the teacher (i.e. the person who creates the exercises), but standard helping goes as follows: when the exercise is a large problem and the learner cannot make a strategy to solve the problem, the exercise is partitioned to smaller problems by subtasking. This way the learner is helped to achieve the final goal by guiding him/her through the smaller tasks. This procedure is in harmony with the idea of easing the learner's mental programming (Korkman 1988).

A concrete example could be 3+2+1. If the learner does not know the answer, the program shows a part of the original problem: 3+2. If the learner can solve that, the original problem can be presented in the form of 5+1. When answering that question correct, the learner has completed also the original problem. The original problem is shown on the screen all the time and the part to be solved is highlighted, so the learner knows the context entirely.

The way of helping is not fixed, so the teacher creating the exercises can decide how the helping path proceeds. Also, the helping is not necessarily a partitioning of the problem. For example, sometimes it is easier for someone to solve addition by adding smaller number to the larger number (i.e. 4+1 is easier than 1+4). With minor modifications to the program code the helping path could be also something else, like verbal help.



| Top |

4. Interdisciplinary design approach

Traditionally, the role of a computer scientist when creating computer-aided education has been that of an implementor. Sometimes the implementing has been disguised as designing, but nevertheless the computer scientists have not been heard during the process. We have been using a novel and interdisciplinary development process to achieve a computer-aided learning environment described in the previous section. A new feature in our collaboration was the role of computer scientists. We, as computer scientists, were not offered a detailed implementation plan or guidelines. It was our task to try to figure out the needs of special teachers and handicapped children, and the potential of computer to meet these needs. Our scheme extends the interdisciplinary dialogue approach of Meisalo et al. (1997) to the multiprofessional context of software development.

 

4.1. Ideal development process

Not seldom, the development of educational software is divided into two phases. In the design phase, a group of educational experts (possibly consisting also of neuropsychologists, cognitive scientists, and occupational therapists) provides a group of computer programmers with a detailed specification of an educational software product. The second phase is the implementation of the software according to the first group's specifications. The requirements come intact from the educational experts, and the programmers are not allowed to take part into the creative process with the first group (Fig. 1). Evaluation takes place only after the programmers have finished their task and the software is already in use.

The two-stage pipeline development scheme of educational software.

Figure 1: The two-stage pipeline development scheme of educational software.

 

In our development process (Fig. 2) there were three improvements to the traditional process of Fig. 1. First, we considered users as an active party in the design. Second, computer scientists were contributing to the design process as opposed to "passive" programmers. The third and most meaningful difference was the pulsating nature of the process; it was not a sequential pipeline of Fig. 1 but a constantly improving and refining design where the flow of information was not restricted to one direction (Fig. 2). At the technical level, the character of the pulsating development process results in flexible data structures which support different perspectives on learning, presented by the conversating partners of the multiprofessional development team. The design of these flexible data structures requires the expertise of a computer scientist, as opposed to an implementation-oriented programmer of the pipeline development scheme.

Improved development process with more participants and more communication.

Figure 2: The pulsating development scheme of educational software in a multiprofessional team.

 

4.2. Finding appropriate data structures to support flexible learning process

Some essential factors in human learning are attentiveness, finding the right problem-solving strategy and upholding the motivation, as well as the amount of units a person can hold in his/her short-term memory (Korkman 1988). As computer scientists, we started from this basis our joint effort with neuropsychologists, special teachers, learners, computer scientists and programmers. Computer scientists' task was to design appropriate data structures to store and represent these factors in a form a computer can understand and act upon (Fig. 2). The data structures should not be static but flexible and dynamic, if the correspondence between human learning and the computer-stored information is to be successfully exploited.

An example of the fruitful interdisciplinary co-operation leading to an appropriate and flexible data structure was the moment of clarification of the individualized learning path. We knew that the learners will be vastly different and unique. Therefore, a personal learning trail was the natural goal in our design. Another idea was to construct the exercises as a cube, where every learner could navigate along his/her optimal learning path through three dimensions, like that of the difficulty level (Fig. 3).

Individualized learning trail through the cube.

Figure 3: Individualized learning trail through the cube.

 

Another example of flexible data structures is the pedagogics of the program. The flow of exercises is controlled by a structure called {\it decision tree}. The decision tree is easily replaceable, since it is a simple text file. In other words, the learning model is independent from the implementation. Therefore, our implementation could be biased to act according to different learning models.

 

4.3. The value of formative evaluation

Essential to the succession of the project was the formative evaluation of the software under development. The idea was to constantly refine the design and the underlying concepts based on the feedback from the educational experts and the observation of the learners using computers. Formative evaluation is at the same time critical and constructive; it is used for finding the problems and to solve them. An important factor in the evaluation was the fact that the evaluation took place during the whole process so that the problems were solved instantly.

In our project, formative evaluation was carried out in frequent get-togethers within the multiprofessional team. In general, however, this kind of formative evaluation in educational software development is still a somewhat uncommon practice. A reason could be the difference in thinking between various disciplines. Every discipline has its own paradigm and prejudices against other disciplines. The means to do science are far from each other, making communication difficult. Sometimes same words are used to refer to totally different concepts.



| Top |

5. The potential of the concept

Computer applications have been developed rapidly and spread to practically every sector of human life. In order to use the latest and greatest computers in education - in a functional way - both computer scientists and experts of learning are needed. The key factor for success is cooperation from the very first stages of planning and development.

Our experiences from multiprofessional software development are promising. In our project, the development scheme resulted in the described data structures promoting an adaptive learning environment. Also the isolation between the actual program and the exercises and their pedagogics turned out beneficial. The ability to change the learned topic and decision-making data modules "on the fly" enabling flexible testing of different learning strategies is handy while evaluating best approaches for a certain situation. We are quite certain that the adaptive approach is scalable to different kinds of education.

Our experiences have encouraged us to continue with the chosen path. The design has been refined and generalized. We are currently working on more universal exercise language, with which different types of exercises can be imported to the program.



| Top |

6. Acknowledgements

We would like to thank neuropsychologist Tuula Eriksson and special teacher Erkki Lamminranta for helping and commenting on the work. We are also grateful to Mikko Koskenniemi, Sami Laasanen and Tuukka Vartiainen for their contribution in the implementation.



| Top |

References

Eriksson, T., Kurhila, J., Sutinen, E. (1997). An Agent-Based Framework for Special Needs Education. Proceedings of the Sixth Scandinavian Conference on Artificial Intelligence SCAI '97, Helsinki, Finland, 270-271.

Korkman, M. (1988). NEPSY. A proposed neuropsychological test battery for young developmentally disabled children. Theory and evaluation. Academic Dissertation, Helsinki, Finland.

Meisalo, V., Sutinen, E., Tarhio, J. (1997). Impacts of interdisciplinary dialogue to Computer Science education. To appear in Proc. IFIP WG 3.2. Working Conference 1997: Computer Science as a Discipline and in other Disciplines.

Ramstein, C., Martial, O., Dufresne, A., Carignan, M., Chasse, P., Mabilleau, P. (1996). Touching and hearing GUI's: design issues for the PC-access system. Proceedings of the second annual ACM conference on assistive technologies ASSETS '96, ACM SIGCAPH, 2-9.

Welsh, M.C., Pennington, B.F., Groisser, D.B. (1991). A normative-developmental study of executive function: A window on prefrontal function in children. Developmental Neuropsychology, 7, 131-149.



| Top | | TIDE 98 Papers |