DISTANCE LEARNING BY INTELLIGENT TUTORING SYSTEM

Part I: Agent-based architecture for user-centred adaptivity

Antonio Fernández-Caballero, José Manuel Gascueña

Computer Science Reseach Institute of Albacete, University of Castilla-La Mancha, Albacete, Spain

Federico Botella, Enrique Lazcorreta

Operations Research Centre, University Miguel Hernandez of Elche, Elche, Spain

Keywords: Intelligent Tutoring System, Agent system, Arch

itecture, E-learning, E-teaching, Adaptivity

Abstract: Agent technology has been suggested by experts to be a promising approach to fully extend Intelligent

Tutoring Systems (ITS). By using intelligent agents in an ITS architecture it is possible to obtain an

individual tutoring system adaptive to the needs and characteristics of every student. The general

architecture of the ITS proposed is formed by the three components that characterize an ITS – the Student

Model, the Domain Model, and the Education Model. In the Student Model the knowledge that the system

has about the student (profile and interaction with the system) is represented. In the Domain Model the

knowledge about the contents to be taught is stored. Precisely, in this model four autonomous agents – the

Preferences Agent, the Accounting Agent, the Exercises Agent and the Tests Agent - have been defined.

Lastly, the Education Model provides the functionality that the teacher needs. Across this module, the

teacher changes his preferences, gives reinforcement to the students, obtains statistics and consults the

matter.

1 INTRODUCTION

Agent technology has been suggested by experts to

be a promising approach to fully extend Intelligent

Tutoring Systems (ITS). By using intelligent agents

in an ITS architecture it is possible to obtain an

individual tutoring system adapted to the needs and

characteristics of every student (Frigo, Pozzebon &

Bittencourt, 2004). In this article, an agent-based

Intelligent Tutoring System architecture for user-

centred adpativity in e-learning/e-teaching of any

matter is introduced. A detailed description of the

agents which monitor the progress of the students

and propose new tasks is also provided. The ITS

proposed is not tied to any course in particular,

being the only requisite that the course has to be

divided into theory, exercises and tests.

Many learning/teaching computer-based

envi

ronments framed in the form of ITS use agent

technology. For example, Cheikes has developed

GIA (Generic Instructional Architecture), an agent-

based software infrastructure devoted to support

rapid development of ITS applications (Cheikes,

1995). Tang carried out the implementation of a

multi-agent intelligent tutoring system for the

learning of computer programming (Tang & Wu,

2000). Capuano has described ABITS, a highly

reusable Intelligent Tutoring Framework suitable to

several knowledge domains (Capuano, Marsella &

Salerno, 2000). A multi-agent system named

MASPLANG developed for the adaptation of the so-

called teaching support units has been introduced

(Peña, Marzo & de la Rosa, 2002). Hospers et al.

have presented an agent-based ITS for nurse

education (Hospers et al., 2003). And there are many

more approaches in distance learning (e.g., Bello &

Bringsjord, 2003; Mota, Oliveira & Mouta, 2004;

Kinshuk et al., 2001; de Antonio et al., 2003; Dorça,

Lopes & Fernández, 2003; Pesty & Webber, 2004;

Baldoni, Baroglio & Patti, 2004).

An ITS usually also incorporates pedagogical

agents

(animated characters) to do learning more

attractive and effective. For example, there is Adele

for medical education (Shaw et al., 1999), and

AutoTutor for the students to learn the fundamentals

of computer hardware, the operating system, and the

Internet (Person & Graesser, 2000). SONIA is the

animated agent incorporated in MASPLANG. The

Fernández-Caballero A., Manuel Gascueña J., Botella F. and Lazcorreta E. (2005).

DISTANCE LEARNING BY INTELLIGENT TUTORING SYSTEM - Part I: Agent-based architecture for user-centred adaptivity.

In Proceedings of the Seventh International Conference on Enterprise Information Systems, pages 75-82

DOI: 10.5220/0002519500750082

 SciTePress

architecture that we introduce in this article does not

incorporate at present any animated agent.

The layout of the paper is as follows. In section 2

a definition of ITS is provided as its most common

features are introduced. In section 3 we define what

an agent is. In section 4 the aims of our agent-based

ITS are explained. From section 5 on, the ITS

architecture is introduced. Lastly, some conclusions

are provided.

2 DEFINITION OF AN ITS

ITS are programs that possess a wide knowledge on

a certain matter, and their intention is to transmit this

knowledge to the students by means of an interactive

individualized process, trying to emulate the form in

which a tutor or human teacher would guide the

student in his learning process (Millán, Agosta &

Pérez, 1999).

Thus, ITS for sure are systems of knowledge

communication. They can be defined that way

because the principal emphasis in the development

of these systems is to provide them with access to

the representation of the knowledge that the system

tries to communicate to the student.

In an ITS the emphasis is put in the knowledge

(what) to being communicated to the student and not

in the mechanism (how) of communication used to

present the knowledge to the student.

Generally speaking, ITS are characterized for

incorporating three models corresponding to three

knowledge levels (see figure 1). Firstly, there is a

Domain Model where the Knowledge of the Domain

is gathered, that is to say the knowledge of what has

to be taught. A Student Model represents the

Knowledge of the Student, that is to say all things

the student knows on the domain. Finally, there is a

Pedagogical Model where the Knowledge of the

Instructional strategies is described; that is to say,

how to teach the Domain Knowledge.

Figure 1: Components of an ITS

3 DEFINITION OF AN AGENT

There is no universally accepted definition for the

term agent, but there are is a wide range of

perspectives in function of the application domain,

the author, and so on.

Franklin and Graesser state: “An autonomous

agent is a system situated within and a part of an

environment that senses that environment and acts

on it, over time, in pursuit of its own agenda and so

as to effect what it senses in the future.” (Franklin &

Graesser, 1996).

Any agent, in accordance with this definition,

satisfies the four properties as indicated next:

• autonomy: agents operate without the direct

intervention of humans or others, and have

some kind of control over their actions and

internal state;

• social ability: agents interact with other agents

(and possibly humans) via some kind of

agent-communication language; agents

collaborate for the sake of performing tasks;

• reactivity: agents perceive their environment,

(which may be the physical world, a user via a

graphical user interface, a collection of other

agents, the Internet, or perhaps all of these

combined), and respond in a timely fashion to

changes that occur in it; in order to respond

effectively to changes, agents have to know at

each instant their surrounding “world”;

• pro-activeness: agents do not simply act in

response to their environment, they are able to

exhibit goal-directed behaviour by taking the

initiative.

4 OBJECTIVES OF THE AGENT-

BASED ITS

The ITS proposed in this paper creates an

infrastructure for distance learning/teaching of a

matter. In accordande with our experience, and in

order to obtain good results, we propose to

decompose the matter to be taught into theory,

exercises and test questionnaires (see figure 2). The

alumni study each topic of the matter reading theory

first, then making exercises and finally answering to

a test. The system will provide help the students

whenever it will be felt necessary.

ICEIS 2005 - HUMAN-COMPUTER INTERACTION

Figure 2: Decomposition of the matter

The first goal of the ITS proposed is that the

alumni learn more and better, that is to say, the

system has to be able to structure learning matter in

such a way to facilitate learning as much as possible.

One the most desirable characteristic to take into

account in learning is the rhythm the student is able

to learn. Thus, the ITS has to adapt the rhythm in

which it introduces the concepts to the learning

rhythm of each student (for instance, to show more

or less exercises, to show more or less tests, etc.).

Another aspect widely considered in learning theory

is reinforcement by rewarding a correct answer and

penalizing the errors (by means of messages, sounds,

etc.).

The second goal in our environment is to

enhance teaching in the same way as learning. One

of the main problems a professor faces when

teaching is that he does not know the skills of his

alumni. Our proposal leads to conclusions that

“teach how to teach”. Within this objective there is

the need to make the matter more comprehensive for

the overall alumni, but always keeping in mind the

requisites given to the subject.

5 ARCHITECTURE OF THE ITS

The general architecture of our ITS (see figure 3) is

formed by the three components that characterize an

ITS, as explained before – the Student Model, the

Domain Model, and the Education Model. In the

Domain Model four agents have been added to

provide the system of a user-centred adaptivity

capacity.

In the Student Model the knowledge the system

has about the student (profile and interaction with

the system) is represented. The model is composed

of three knowledge databases (KDBs). (1) The

Personal Information KDB stores the necessary

personal information of the student to control his

access to the system. (2) The Profiles KDB stores

the level as well as the presentation styles of the

students. The students are assigned different levels

depending on their learning rhythm. (3) The

Learning KDB stores parameters such as the

exercises and tests proposed so far to the students,

the time spent on answering the questionnaires, the

pages of theory visited and the scrolls performed on

those pages, or the reinforcement material prepared

by the Pedagogic Module.

In the Domain Model the knowledge about the

contents to be taught is stored. This model consists

of four KDBs: (1) the Theory KDB incorporates the

pages of theory that have been prepared for teaching

the matter, (2) the Tests Questionnaire KDB stores

the battery of test questions related to the matter, (3)

the Exercises KDB stores the battery of exercises on

the matter, and, (4) the Reinforcement KDB contains

the information used by the Pedagogic Module to

prepare the material to be shown when a student

needs to be reinforced.

DISTANCE LEARNING BY INTELLIGENT TUTORING SYSTEM. Part I: Agent-based architecture for user-centred

adaptivity

Figure 3: Architecture of the agent-based ITS system

The Pedagogic Module provides the necessary

mechanisms to efficiently present the matter to the

student. This module is in charge of carrying out

three tasks: (1) to provide the learning guidelines for

the student (including any necessary reinforcement

provided by the system), (2) to update statistics in

the Domain Model of the exercises and tests

presented, (3) to store into the Learning KDB

important data such as the material prepared to

reinforce the student who needs it, the responses

given by the student to the exercises and tests

proposed, as well as the scores that the student has

gotten and the time that he has spent in reaching the

aims.

The Preferences Agent supervises the user

preferred style of presentation (type and size of

letter, colors, margins, and so on). When the user

changes his style of presentation the Preferences

Agent creates a personalized sheet of styles for the

user and updates the user's interface in accordance

with his new pleasures. The information that this

agent gathers is stored in the Profiles KDB. The

Accounting Agent observes the student interaction

with the interface when the pupil accesses a page of

theory. When the student changes to another page of

theory, the Accounting Agent stores in the Learning

KDB some valuable information (the name of the

visited page, the time that the student has spent on it

and the scrolls performed on it). The Exercises

Agent takes charge of choosing the exercises that

will be proposed to the student in the topic that he is

currently studying. This agent stores the chosen

exercises in the Learning KDB as well. In the same

way, the Tests Agent is in charge of choosing the

test questions that will compose a test questionnaire

proposed to the student in the topic that he is

studying at this moment. The test questions selected

are also stored in the Learning KDB. The Exercises

Agent and the Tests Agent do the selection when the

student finishes the first visit to the first page of

theory of every topic. We may highlight that the

Exercises Agent and the Tests Agent are proactive

because they carry out their tasks in parallel with the

activity that the student performs. Indeed, the

student is reading theory without realizing the work

of both agents.

Lastly, the Education Model provides the

functionality that the teacher of the system needs.

Across this module the teacher changes his

preferences, gives reinforcement to the students,

obtains statistics and consults the matter. This model

is in fact devoted to help the teacher to change the

contents of the matter on the basis of the information

obtained from the Student Model and the Domain

Model.

ICEIS 2005 - HUMAN-COMPUTER INTERACTION

Figure 4: Agent class diagram

6 DESIGN OF THE AGENTS

As it may be observed in figure 4, agents have been

implemented as applets.

6.1 Preferences Agent

The Preferences Agent supervises the style of

presentation that the user likes. The Preference

Agent perceives the interaction of the student with

the user interface and acts when he changes his

tastes. The preference agent is continually running to

know the student’s preferences at any time.

The p

rocess that follows when the user decides

to change his visual preferences is shown in figure 5

as an activity diagram for activiy “Change

preferences”. When the student decides to “Change

preferences”, the Preferences Agent shows him a

form with the preferences that he has selected up to

this moment. This way the user can perform the

changes when he considers that are appropriate.

After having completed the form, the new

selected preferences are updated and an example

page is shown to the student with all the features of

the new selected style of presentation. If the student

does not like the page, he may continue changing his

preferences.

Figure 5: Activity diagram for “Change preferences”

DISTANCE LEARNING BY INTELLIGENT TUTORING SYSTEM. Part I: Agent-based architecture for user-centred

adaptivity

Figure 6: Activity diagram for “Detection of scroll”

6.2 Accounting Agent

The Accounting Agent perceives the interaction

between the student and the user interface and acts

(gets information) when the student changes to

another page of the ITS, scrolls up and/or down a

page, performs an exercise or a test, and so on.

Let us focus on the Accounting Agent when

watching the interaction of the student with the

interface in theory pages. Here, more concretely, the

agent is in charge of watching the scroll that the

student performs on a page of theory as well as the

time that he has remained in that page. When the

student leaves studying a page of theory, the

Accounting Agent stores all parameters gathered

during this time (scroll and time of permanence) in

the database.

In figure 6 the algorithm to detect the scroll that

the student performs when he visits a page of theory

is shown. Once the student has entered a theory

page, he may advance in his reading or go back in

the page. Whilst the student is advancing through the

page, the value of “Greatest advance” is being

updated. Now, when he steps back the value of

“Greatest backward” is updated. Notice how all

steps are stored in the database as “Scroll History”.

6.3 Exercises Agent

The Exercises Agent is in charge of choosing the

exercises that will be proposed to the student in the

topic that he is currently studying. The Exercises

Agent is autonomous as it controls its proper actions

in some degree. The agent, by its own means (pro-

active), selects the set of exercises to be proposed in

the subject studied by the student and adds to each

exercise the links to the theory pages that explain the

concepts (or topics) related to the exercise. When

solicited, it sends the page containing the exercises

to be proposed.

ICEIS 2005 - HUMAN-COMPUTER INTERACTION

Figure 7: Exercises Agent state diagram

As it may be observed in figure 7, the Exercises

Agent state diagram, when the student has just

visited for the first time the first page of a topic, the

Exercises Agent shows the selection of exercises

that will be proposed to the student for the topic. If

the student is a level-1 student (low level student),

the agent selects the more basic exercises (state

“Elaborate basic exercises”) and later on the more

complex exercises (state “Elaborate complex

exercises”). Now, if the student is a level-2 student

(high level student), the agent is only allowed to

select the complex exercises. Once the agent has

selected the exercises it will remain inactive (in an

“Idle” state) while the student does not go on to the

following topic.

6.4 Tests Agent

Similarly, the Tests Agent is in charge of choosing

the test questionnaires that will compose the test that

will be proposed to the student in the topic that he is

studying. The Tests Agent is also waiting until it is

asked for tests questionnaires pages. The agent by its

own means (pro-active) goes on designing a set of

tests for the subject the student is engaged in.

Figure 8: Tests Agent state diagram

As you may observe in figure 8 – the Tests

Agent state diagram -, the Tests Agent performs the

selection of test questionnaires at the same time that

the Exercises Agent performs the selection of

exercises. Once it has selected the test

questionnaires, the agent will remain inactive (“Idle”

state), while the student does not go on to the next

topic.

7 CONCLUSIONS

In this paper we have proposed an architecture that

considers the high diversity of users’ skills and

preferences: a user-centred and adaptive interaction

multi-agent system. Our model proposed has been

applied to e-learning/e-teaching by taking advantage

of the current state of the art of ITS. A way to insert

user adaptivity into an ITS is by using agent

technology. This is due to the characteristics that

DISTANCE LEARNING BY INTELLIGENT TUTORING SYSTEM. Part I: Agent-based architecture for user-centred

adaptivity

intelligent agents possess – autonomy, social ability,

reactivity and pro-activity. I this article, we have

introduced an agent-based ITS architecture that

enables a better learning to the students and a better

teaching to the professors.

In this sense, in our distance learning system we

have introduced a Student Model, a Domain Model,

and an Education Model. In this latter model four

agents – the Preferences Agent, the Accounting

Agent, the Exercises Agent and the Tests Agent -

have been proposed. To conclude, the multi-agent

system described in the paper gets data obtained

from the profiles to adequate the contents shown to

the concrete student that accesses the distance

learning ITS. On the other hand, the multi-agent

system obtains measures that permit to get

recommendations to enhance the course. This way,

jointly e-learning and e-teaching are greatly

enhanced.

ACKNOWLEDGEMENTS

This work is supported in part by the Spanish Junta

de Comunidades de Castilla-La Mancha PBC-03-

003 and the Spanish CICYT TIN2004-08000-C03-

01 grants.

REFERENCES

Baldoni, M., Baroglio, C., Patti, V., 2004. Web-based

adaptive tutoring: an approach based on logic agents

and reasoning about actions. Artificial Intelligence

Review, 22(1):3-39.

Bello, P., Bringsjord, S., 2003. HILBERT & PATRIC:

Hybrid intelligent agent technology for teaching

context-independent reasoning. Educational

Technology & Society, 6(3):30-42.

Capuano, N., Marsella, M., Salerno, S., 2000. ABITS: An

agent based Intelligent Tutoring System for distance

learning. Proceedings of the International Workshop

on Adaptive and Intelligent Web-Based Education

Systems, ITS 2000.

Cheikes, B.A., 1995. GIA: An agent-based architecture for

Intelligent Tutoring Systems. Proceedings of the

CIKM'95 Workshop on Intelligent Information

Agents.

de Antonio, A., Imbert, R., Ramirez, J., Mendez, G., 2003.

An agent-based architecture for the development of

intelligent virtual training environments. Second

International Conference on Multimedia and ICTs in

Education, m-ICTE 2003.

Dorça, F.A., Lopes, C.R., Fernández. M.A., 2003. A

multiagent architecture for distance education systems.

Proceedings of the 3rd IEEE International Conference

on Advanced Learning Technologies, ICALT’03, page

368.

Franklin, S., Graesser, A., 1996. Is it an Agent, or Just a

Program?: A Taxonomy for Autonomous Agents.

Intelligent Agents III, Agent Theories, Architectures,

and Languages, ECAI '96 Workshop (ATAL), Lecture

Notes in Computer Science 1193:21-35.

Frigo, L.B., Pozzebon, E., Bittencourt, G., 2004. O papel

dos agentes inteligentes nos sistemas tutores

inteligentes. World Congress on Engineering and

Technology Education, page 86.

Hospers, M., Kroezen, E., Nijholt, A., op den Akker,

H.J.A., Heylen, D., 2003. An agent-based intelligent

tutoring system for nurse education. In Applications of

Intelligent Agents in Health Care, J. Nealon and A.

Moreno (eds), pages 143-159.

Kinshuk, Han, B., Hong, H., Patel, A., 2001. Student

adaptivity in TILE: A client-server approach.

Proceedings of IEEE International Conference on

Advanced Learning Technologies, ICALT2001, pages

297-300.

Millán, E., Agosta, J.M, Pérez J.L., 1999. Application of

bayesian networks to student modelling. Proceedings

of PEG'99: Intelligent Computer and

Communications Technology: Teaching & Learning

for the 21st Century.

Mota, D., Oliveira, E., Mouta, F., 2004. MyClass: A Web-

based system to support interactive learning in virtual

environments. Workshop on Modelling Human

Teaching. Tactics and Strategies.

Peña, C.I., Marzo, J.L., de la Rosa, J.L., 2002. Intelligent

agents in a teaching and learning environment on the

Web. Proceedings of the 2nd IEEE International

Conference on Advanced Learning Technologies,

ICALT2002.

Person, N.K., Graesser, A.C., and the Tutoring Research

Group, 2000. Designing AutoTutor to be an effective

conversational partner. Fourth International

Conference of the Learning Sciences, pages 246-253.

Shaw, E., Ganeshan, R., Johnson, W., Millar, D., 1999.

Building a case for agent-assisted learning as a

catalyst for curriculum reform in medical education.

Proceedings of the International Conference on

Artificial Intelligence in Education, pages 509-516.

Tang, T.Y., Wu, A., 2000. The implementation of a multi-

agent intelligent tutoring system for the learning of

computer programming. Proceedings of 16th IFIP

World Computer Congress-International Conference

on Educational Uses of Communication and

Information Technology, ICEUT 2000.

Pesty, S., Webber, C., 2004. The Baghera multiagent

learning environment: an educational community of

artificial and human agents. Upgrade, Journal of

CEPIS (Council of European Professional Informatics

Societies), 4:40-44.

ICEIS 2005 - HUMAN-COMPUTER INTERACTION