Educational Technology & Society 1(1) 1998
ISSN 1436-4522

The Necessity and an Outline of an Integral Approach to Designing Intelligent Tutoring Systems of a New Generation

Towards Increasing Effectiveness of Knowledge Transfer on the Basis of Ideas and Methods of Artificial Intelligence and Cognitive Science

Moderators & Summarizers:
Vladimir A. Fomichov
Professor of Computer Science and Cognitive Science, Department of Information Technologies, K.E.Tsiolkovsky Russian State Technological University, Orshanskaya str. 3, 121552 Moscow, Russia. Email: vaf@nw.math.msu.su

Olga S. Fomichova
Head of the Studio "Culture, Ecology, and Foreign Languages", Moscow Children and Teenagers Palace for Creative Work, Kossygin street 17, Moscow, Russia. Email: vladfom@yahoo.com


Discussion Schedule
Discussion: 1-9 August 98
Summing up: 10-12 August 98


Pre-discussion paper

1. Shortcomings of the Dominant Current Approach to the Design of Intelligent Tutoring Systems

The XXth century is moving to its ending. During the second half of this century, science and technology have received a lot of outstanding results. However, many millions of children throughout the world encounter huge difficulties in studying various school disciplines: from mathematics and physics to literature and foreign languages. The inability to master school disciplines is a very essential (probably, principal) cause of numerous conflicts between children and adults – teachers and parents. The shortcomings of school education are the cause of difficulties encountered by many college students in mastering the learned disciplines.

In the beginning of the 1980s, many specialists in Computer Science (especially, in Artificial Intelligence) believed that the design of intelligent tutoring systems (ITSs) may considerably contribute to the progress in preuniversity and university education. Many efforts have been undertaken in this direction. However, the accumulated experience has not confirmed these hopes. In general, the use of ITSs has not improved the situation as a whole.

It appears that the principal shortcomings of the major part of the existing ITSs are as follows.

1. Usually, students must select one correct answer from several suggested variants. We believe that this approach has serious drawbacks even as concerns the learning of mathematics and physics. And we agree with Prof. K. De Smedt (de Smedt, 1998, URL:http://www.futurehum.uib.no) that this approach is of a very little use for studying the humanities, because, in fact, it doesn't develop the creativity of the student, doesn't enable the student to express his/her personality.

2. ITSs destined for the learning of special disciplines are very far from being so thrilling for the students, carrying them away so much as numerous computer games.

3. One constructs ITSs not taking (or taking only a little) into account the role of the teacher in the learning process.

4. The existing ITSs are destined for solving particular tasks of education and don't form one or several large families solving together a complex of interrelated tasks of education and up-bringing.

5. The existing ITSs are destined mainly for studying particular disciplines and are not oriented at developing the personality of the child or the university student as a whole, at transferring the knowledge about how to be successful in the life (in particular, how to understand people, how to communicate with people).


2. What is Necessary

We suppose that it is necessary to elaborate an up-to-date paradigm of constructing ITSs of a new generation. To our opinion, the ideas of Artificial Intelligence theory, Cognitive Science (including our Theory of Dynamic Conceptual Mappings), and Philosophy of Language do provide such a possibility.

The focus of the discussion below is basic principles of designing ITSs destined for helping to achieve the needed level of knowledge by the most part of different age groups of the society. I.e., the focus is the ITSs destined for the use in the classroom. However, the formulated principles may be easy used and seem to be precious for constructing ITSs destined for the individual work at home.

We see the basic principles of constructing intelligent tutoring systems of a new generation as follows.

1. The accent in the design of ITSs is to be moved to constructing systems stimulating the learners to carry out some activities of exploratory character, developing the creativity of the learners.

An important feature of such systems is that the learner doesn't select one correct answer from several suggested variants but submits a result of some creative activity – most often, a natural language text (a story, a poem, a solution of a mathematical or physical task, etc.) or a figure.

An essential part of such ITSs is to enable the learner to express his/her personality – for instance, in figures, stories, etc.

2. ITSs of a new generation must carry away the learners even more than the various computer games not destined for realizing educational purposes.

However, CONTRARY to the case of computer games, ITSs must NOT EXCITE the learners (especially, children); in particular, must not cause an outburst of laugh at lessons.

This demand may be satisfied in case the designers of an ITS proceed from the COMPLETE inner world's picture, or conceptual system, of the learners and manage to "switch on" the imagination, the intelligence of the learners into the processes of acquiring knowledge. The developers of the scripts for the ITSs must "inscribe" very exactly and carefully the new pieces of theoretical materials into the semantic traces made in the conceptual system of the learner by the previous experience of every-day life and emotional experience (under the influence of, e.g., seen pictures of the nature, read books, seen films, etc.).

3. ITSs of a new generation must become an organic part of the teaching process, including the teacher. This means, first of all, the following.

Suppose that some ITS is destined for the learning of some piece of theoretical materials. Then the work with such a system must consist of such forms of activities and cause such an emotional state of the learners which HARMONIZE quite well with the GENERAL ATMOSPHERE in the classroom at the lessons including the learning of this theoretical subject.

On one hand, in this case the teacher may alternate his/her oral explanations with the use of ITSs by the students. On the other hand, practically always in the classroom there are students who objectively demand a particular attention of the teacher, a personal communication with the teacher during a lesson. Some of these students may be very gifted, the other may encounter some difficulties in studying just this piece of materials (in particular, after an illness, etc.).

When there is an ITS for explaining the considered piece of theoretical materials, and this ITS harmonizes quite well with the general atmosphere in the classroom at the corresponding lesson, then the teacher has the possibility to ask the most part of the students to work with such an ITS for acquiring this piece of materials. Due to this, the teacher can personally communicate with one or several students objectively requiring his/her personal attention just at this particular lesson.

The existence of such a possibility for the teacher at every lesson is a very important condition of attaining the needed level of knowledge by all students.

But when an ITS EXCITES the learners (in particular, causes an outburst of laugh) then the consequences for the teaching process are unforeseen. Most often, students (especially, children) will be unable to return to the emotional state being necessary for successful learning activity during the remaining part of the lesson.

This explains additionally, why the ITSs must carry away the learners more than computer games but must NOT EXCITE the learners.

4. ITSs must give the teacher the possibility to select the most appropriate kinds of the learning activity proceeding from the estimation of the emotional state of the learners. For instance, it would be desirable to suggest the different forms of the work with the ITS in the humanities after a soccer match and after a voluminous test in mathematics or physics.

5. The designers of ITSs of a new generation must proceed from the positive experience of knowledge transfer obtained by the teachers (in a broad sense: school teachers, university instructors, etc.) of various disciplines.

6. The dominant part of knowledge in the society is stored and conveyed by means of natural language (NL). The ability to find the most appropriate language constructions for expressing one's own opinion, the arguments, questions, objections, etc. is highly important for the success in life.

It is known that in many countries with different level of economic development, very many children are unable to master such sublanguage of their mother tongue which is necessary for building a successful carrier. Meanwhile, there are scientific and experimental data showing how positively the development of children's language abilities influences the development of their thinking and mental outlook.

That is why a considerable accent in creating the ITSs of a new generation must be made on developing the ability of young children and teenagers: (a) to understand the written language, to reconstruct the complete situations mentioned in a text, to see the macrostructure of a discourse; (b) to understand a partner in a dialogue; to choose the most appropriate constructions for expressing the thoughts; to conduct a dialogue.

7. It follows from the principles 1 and 6 that, very often, the students will submit the fulfilled tasks in the form of NL-texts.
That is why the progress in the creation of ITSs of a new generation will depend in many cases on the use and adaptation of information technologies developed in the field of NL processing.

It seems important to note that the use of linguistic information technologies is important for constructing one part (though considerable) of the ITSs of a new generation. And the other part may be elaborated with the help of more traditional technologies.

8. ITSs of a new generation must be oriented (as a whole) not only at studying various special disciplines (mathematics, biology, foreign languages, etc.) but also at transferring the life experience of adults, the knowledge how to be successful in the life (in particular, how to understand the other people, how to take into account the emotional state, intentions, and beliefs of a dialogue partner, etc.).

Such ITSs are to contribute to diminishing the gap between the generation of adults and the generation of children.

9. The recent studies in the Philosophy of Language have shown that social reality is constructed by means of symbols, and the main stock of such symbols is formed by natural language expressions (see, e.g., John Searle, The Construction of Social Reality. Free Press, New York, 1995). Numerous conventions taking place in the society are formulated by means of NL.

Hence for successful social adaptation of children (including teenagers), it is highly important to develop properly their abilities of effective processing symbolic information.

In this connection, very large prospects for the designers of ITSs are opened by a unified, informational-cybernetic approach to teaching natural language (mother tongue, foreign languages) and the language of painting.

The role of symbols in NL-texts is played by words and word combinations. In accordance with the ideas of modern Semiotics, pictures may be considered as texts whose symbols are combinations of lines, colours, etc. There are scientific and experimental data showing a high educational potential of such an approach to teaching the language of poetry and the language of painting when metaphors are interpreted as symbolic expressions conveying the thoughts and emotions of a poet, and pictures are considered as the encoded messages sent (may be, several centuries ago) by their authors – painters.

ITSs of a new generation realizing a unified approach to teaching languages, literature, poetry, art from the standpoint of symbolic information processing will highly extend the life space of the learners, make easier the learning of painting and music, and will greatly contribute to successful social adaptation of children and young adults.

10. The ITSs of a new generation must constitute one family or several large families destined for solving together a complex of tasks of education and up-bringing (in case of children and young adults – in particular, university students).

For this, it is necessary to distinguish a system of such features and abilities of the personality that each concrete ITS from a family contributes to developing one or several such features and abilities (and, in most cases, helps also to transfer some part of knowledge pertaining to special disciplines).

Nowadays, the existence of a great number of various computer educational tools makes it difficult to use them in the classroom and very difficult to include them into the processes of teaching and up-bringing.

If we decide to elaborate and realize a new paradigm outlined above, every computer program will find its place on the branches of this paradigm. In this case, the whole picture can be compared with the fireworks: a great number of intelligent tutoring systems will make a sparkling picture with its own idea clear to the users: teachers, students, and parents.

3. Goals of the Discussion

With respect to the above-said, we initiate the discussion with the following goals:
  • to find out the opinions of educators and specialists elaborating educational technologies about creating such an integral approach to designing ITSs of a new generation;
  • to distinguish a circle of specialists interested in creating such an integral approach and in implementing the stated principles;
  • to explicate and generalize the accumulated experience in education which may become a starting point (points) for elaborating a new paradigm for constructing ITSs;
  • to explicate the achievements in the educational technology and in formal methods which may underlie constructing ITSs of a new generation.

Post-discussion summary

The discussion showed the necessity to think about a Teacher Model analysing different teaching styles (Ashok Patel) and even about a more general CONSISTENT MODEL of an education system/process in the future (Rossen Rashev). These ideas seem to be more promising than having a try to categorize the learners by learning styles.

Dennis R. Nelson suggested the interesting idea of believing that “the true teacher model is the parent”. We suppose that the parents and the teachers have different rights as concerns a child. That is why, to our opinion, this analogy can’t be accepted as a whole. However, it enables us to put forward the following constructive hypothesis: analysing the successful methods of teaching and up-bringing children, we have the chances to find a lot of ideas which may prove to be effective as concerns the transfer of knowledge to the grownups.

Greg Hume, Cossi Grandguenon, and Angel de Vicente supported the idea of modeling the emotional state of the student. It is worth mentioning that the emotional state of a student at the moment of acquiring knowledge can influence the result of it greatly.

The admission of the existence of a gap between entertainment and education caused a discussion about learning and playing. John Eklund stated that he did not make the distinction between play and learning: "“Play is learning. Learning need not be a serious activity at all”. This viewpoint (supported by Mike Peay, David Graves, and some other participants of the discussion) led to the question whether the very act of “consciously learning” caused additional stress (Ashok Patel).

In response to it, Akihiro Kashihara agreed that it was true in case the learners had no/little motivation to learn. Meanwhile, he stressed that the key point of learning in this view was not conscious/unconscious but the extent of motivation to learn.

It is clear that motivation would come from different things: curiosity, necessity, obligations, fun, etc. In order to extend the motivation to learn, one should extend, enrich the inner conceptual picture of the world of a child, making it possible for him/her to enjoy the processes of reasoning, reflection, symbolic information processing , etc.

A special attention should be paid to creativity as one of the main feature of Intelligent Tutoring Systems (ITSs). The idea of constructing systems stimulating the learners to carry out some activities of exploratory character, developing the creativity of the learners was supported by Clark N. Quinn. Gary B. Grant stressed that computers offer interactivity and control which should promote creativity.

The idea of “real life” learning experience was supported by William R. Terrell. It proves to be very useful while teaching students, postgraduates. As for teaching kids, we also should take into account their life and emotional experience.

It should be stressed that we should be very careful deciding whether to design ITSs for kids or not (Vilnis Vulfs, John Schwartz, Gary B. Grant). Certainly, kids do not need to get lost in some virtual world, because they are happy to penetrate the real world surrounding them.

Though modern ITSs have a number of drawbacks mentioned during the discussion, and they are very far from being thrilling, it is possible to elaborate a new conception for designing ITSs of a new generation. Henry Brzeski stated that many students needed to see a human face. Clark N. Quinn supported the idea and preferred to have teachers in the loop. And it should be one of the main demands to ITSs of a new generation.

Katy Luchini supported that ITSs have a great potential for bridging a gap between entertainment and education. Starting bridging this gap, all the participants of the discussion came to the conclusion that “to laugh is better than not to laugh”, because “laughing brings and gives us real motivation of learning” (Arun-Kumar Tripathi), but we should be careful lest we should have a mess-up situation. By “laughter” we are meaning engagement and interactivity (John Eklund), and a teacher should certainly joke and make children laugh.

It is very important that our discussion attracted the attention of practical teachers (Cossi Grandguenon, Natalie Perzylo). Because ITSs are destined to be successful with teachers in the loop.

Kelly Tribble noted that “all forms of play are learning … but not all forms of learning are play. Perhaps we should …”.

Why not, if we find the right definition for playing, right for kids and right for adults.

Mike Collett and Tom Abeles saw in the materials of the formal discussion the tendency to consider ITSs as a reinforcement of existing methods of teaching rather than contributing to a new generation.

We suppose that there is no contradiction between (a) the aim to proceed from the best methods and experience of face-to-face-teaching in the classroom while elaborating new ITSs and (b) the aim of constructing ITSs of a new generation.

Our own interests are mainly in teaching children. There are known various approaches to teaching children, and we do know for sure that some of these approaches are very effective, flexible and contribute to forming a highly positive attitude of the child to the process of acquiring knowledge.

These methods by far exceed the possibilities of existing ITSs. Hence the task of constructing new ITSs with respect to the best accumulated theoretical experience and practical experience of teaching is a progressive one.

Obviously, in case technical achievements provide the possibility to realize some new effective manners of teaching and learning, it would be natural to try to do it and to analyse the results.

In this connection, we would like to note that one should be very careful while realizing new technical possibilities in ITSs or in distance education.

For instance, Mike Collett supposes that the learners must be able to add content to the system to be used (or not) by others. To our opinion, there are such effective, subtle, exact methods of explaining to children difficult pieces of theoretical materials (for instance, some pieces of a foreign language grammar) that the mentioned technical ability will contribute to inserting unuseful content into the system, to stimulating unproductive associations, questions, ideas and to diminishing the effectiveness of the learning process.

On the other hand, we share the ideas of Mike Collett about the importance of taking into account the peculiarities of the culture, language, and educational needs of different users throughout the world and about the necessity of elaborating new economical foundations of constructing and using ITSs and using distance education.


decoration