OBSERVATION AND SUPPORT IN A COLLABORATIVE

PEDAGOGICAL DEVICE

A Multi-agent System to Control and Monitor Pedagogical Activities

Insaf Tnazefti-Kerkeni

1,2

, Henda Belaid-Ajroud

and Bénédicte Talon

1,2

Univ Lille Nord de France, F-59000 Lille, France

LISIC, ULCO, F-62228 Calais, France

URPAH, Faculté des Sciences de Tunis, Le Belvédère, Tunisie

Keywords: Collaborative Learning Environment, Monitoring of Activities, Multi-agents Systems, e-Learn 2.0.

Abstract: This article discusses the agentification of coordination and control of a collaborative pedagogical device. It

presents the advantage of a multi-agent system to instrument some features of a collaborative educational

device and describes its implementation in the context of a pedagogical engineering studio. The article

interests at first in the context of insertion of the device: why and how the MAETIC device is proposed to

work out vocational skills training with ICT. The article then describes the function of the MACCADAM

studio and the specific devices it generates. It then presents the benefits of a multi-agent system in this

context and described the monitoring system proposed for the management of the actors. The article ends on

the progress of work and the prospects so far.

1 INTRODUCTION

Teachers must be trained and/or assisted in the

educational use they make of technologies.

We work to produce a Pedagogical Engineering

Studio (PES) called MACADDAM (Talon and

Leclet, 2011). The PES assists teachers in their

efforts to design educational devices. Generated

devices are dedicated to education through

collaborative projects. Devices are designed with

available and easily exploitable tools in order to

meet the needs of the teachers.

ELearning 2.0 (Downes, 2005) allows teachers to

exploit Web 2.0 applications to construct

educational devices. An eLearning 2.0 device is then

based on a combination of specific features of Web

2.0 applications: Forum (discussion, negotiation,

argument), Wiki (collaborative writing), documents

management (sharing, storage), Blog (Editorial and

comment functionalities, document storage and

sharing), etc. The advantage of this type of

construction is that Web 2.0 tools are directly

available, are generally free and can easily be used

in an educational setting (Williams and Jacobs,

2004) (Caron, 2007). These environments are under

the complete control of the students and/or teachers

who can promote their products. The actors in these

environments may be released from the

administration constraints generally associated with

traditional architectures and easily share access to

their resources.

However, in the current version, services offered

by the generated pedagogical devices have a lack of

monitoring. There is no management of interactions

taking place within the educational system.

However, in the field of collaborative learning,

management of traces is important because it is

necessary to analyze information about actors and

their activities (Settouti, 2010). It provides the

trainer accurate and adequate information to track

individual and collective participation. This allows

to conduct the evaluation of the devices and favors

their re-engineering.

We describe here a multi-agents system (MAS)

(Ferber, 1999) to be coupled to the MACADDAM

studio. This system allows implementing

functionalities to keep a detailed history of actions

of students and student groups performed on the

eLearning 2.0 platform. This archiving allows

analyzing the life of every individual, group and its

evolution.

135

Tnazefti-Kerkeni I., Belaid-Ajroud H. and Talon B..

OBSERVATION AND SUPPORT IN A COLLABORATIVE PEDAGOGICAL DEVICE - A Multi-agent System to Control and Monitor Pedagogical Activities.

DOI: 10.5220/0003904001350140

In Proceedings of the 4th International Conference on Computer Supported Education (CSEDU-2012), pages 135-140

ISBN: 978-989-8565-06-8

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

2 THE MACADDAM STUDIO

2.1 MAETIC Devices

In order to work out domain and transverse skills,

teachers have designed educational devices using

Web 2.0 tools. A study of these devices was

managed to extract a pedagogical method called

MAETIC. MAETIC is dedicated to the management

of project-based pedagogy in-group. It was validated

through successive evaluations (Leclet and Talon,

2008). MAETIC aims at developing professional

skills and guides groups of students in all stages of

the project. The part of the system dedicated to

students is called "MAETIC e-suitcase". The e-

suitcase includes an access to the teacher’s logbook

(important information, activities of the session,

etc.), an access to teacher’s resources (course

materials, exercises to do, etc.) and an access log of

the group of students. That part of the pedagogical

device dedicated to the teacher is called "MAETIC

Toolbox". A toolbox provides mechanisms to feed

the teacher’s logbook, to check students' logbooks,

to comment on their work, to assess their work, etc.

MAETIC method is based on an in-group

project-based pedagogy. The student constructs

knowledge through projects and formulates his/her

own problems (Schneider and al., 2005). The work

interaction is the cause of socio-cognitive

confrontations that have positive effects on the

representation of the task, on the goals to be

achieved and on the procedures to achieve them as

well as the control of cognitive activities (Roux,

2004). Figure 1 shows the interactions between users

and the elements of the device and the activities

performed by each user of the system. Each of actors

uses resources, handles tools and implements

activities.

So, in a MAETIC device, the teacher:

• Uploads articles and resources on his/her

logbook. This logbook informs students about the

life of the teaching unit. Posts provide general

comments on the work and on its progress.

• Handles tools that enable him/her to

communicate with students.

• Oversees the work of the groups. Thus, he/she

can view, download and comment on the activities

of the groups via the student groups’ logbook.

The student:

• Consults or download resources made available

and accessible via the teacher’s logbook. He must

consult the teacher’s logbook before each session.

• Uses tools that allow him to communicate with

other students or with the teacher.

• Realizes activities related to the planned project.

The teacher helps to define these activities. The

report and implementation of these activities are

recorded on the group’s logbook.

Once the group is defined, students will choose

among themselves a student who will assume the

task to manage their logbook. This logbook, aims to

keep the teacher regularly informed on the progress

of the project. It is used to describe the life of the

project, making available the developed

deliverables, the report of the activities and

information on the project.

2.2 The MACCADAM Project

In order to assist the teacher in the production of

MAETIC devices, the MACADDAM project is

developing a PES. The PES helps teachers to deploy

their own educational devices dedicated to the

implementation of MAETIC environments. The

studio assists them in the formulation of needs

(design support) and releases them of tasks

associated with the deployment a teaching device

(development aid). The devices are designed and

instrumented using Web 2.0 tools.

The studio controls the design process. The PES

allows the selection of the elements that enable to

Figure 1: User Interactions in a MAETIC device.

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

136

define the device best suited to educational uses. The

PES assists the teacher in the formation of his/her

pedagogical script and generates the e-suitcases and

the toolbox. The studio offers the ability to select

and identify activities to implement in the script.

However, resources and activities appropriated to

the training area are under his/her responsibility.

2.3 PES Needs and Contribution of

Multi-agent Systems

Communication between users (teachers and

students) in a MAETIC device is mainly done via a

Weblog. Weblog technology provides a very

consistent material, easy to collect but more difficult

to analyze. The time spent by the teacher to monitor

and analyze the activities of the student is higher

than the time spent in the traditional classroom (Al-

Sakran and Serguievskaia, 2008).

A statistical analysis of interaction justifies the

adding of automatic assistance that should provide

MAETIC. The assistance includes tools for the

teachers in charge of managing groups.

The statistics about interaction enable to evaluate

the group’s life and its evolution. The status

indicators of progress and sustainability of the group

especially interest us. Among indicators, one can

cite:

• the identification of work overload for a given

student so he/she can be exempted to perform some

activities,

• the possibility of extending or shorting

completion deadlines of an activity (change the

training schedule),

• the assessment of the state of completion of an

activity or a task,

• the evaluation of social relationships and

productivity of a student, and so on.

The aim is to help students in the realization of the

project and in collaborative learning and also to

assist the teacher in the monitoring of both

individual and collective activities.

The assistance issues need to collect traces of

activities performed by the students. Therefore, we

are developing a system which collects and analyzes

data from the project activities. This system must

trace the activity of the actors in order to make the

most meaningful analysis. It must analyze the use of

the environment and the data generated in this

environment (forum, mail, meetings, etc.).

We have opted for an incremental and iterative

approach for the design of the observation and

assistance system. The devices generated by the

MACADDAM studio will be equipped with this

system. It is based on a multi-agent architecture

described in the next section.

The choice of a multi-agent architecture for the

observation and assistance-type system is motivated

by several reasons:

1. From a programming point of view. it is possible

to add new agents or modify the behavior of existing

agents without affecting the overall structure. This

allows an iterative and incremental development.

2. We are interested by the ability to solve

distributed problems in a multi-agent architecture.

We propose to identify agents that are specialized in

observation tasks and others that are specialized in

assistance tasks.

3. We are faced to a distributed environment. The

multi-agent approach offers distributed agents being

able to communicate.

3 PROPOSED ARCHITECTURE

3.1 Analysis of the System

We are interested to provide observation and support

tools to ensure the following functions:

• Course management (self-diagnostic tools,

course management software, etc.);

• Review and validation (assistance to individual

and collective reviewing, etc.).

We have identified three spaces in the

MACCADAM studio: a teacher’s space, a student’s

space and a group of students’ space. Each space has

a descriptive name, functionalities, educational

resources, technological tools (Web 2.0 tools) and

functional tools for the observation of use.

The studio generates the tools at the beginning of

a learning session:

• Technological tools are based on web 2.0

technologies. They are tools that the actors need to

perform activities in their space. These include, for

example, the student’s logbook and the teacher’s

toolbox.

• Functional agents are tools for the observation of

use. These tools aim to mark out the behaviour of

students, groups of students and teachers. They aim

too to analyze traces.

After an evaluation process, we chose MaSE

methodology (Multi-agent System Engineering)

given by (Deloach, 2004) for the development of our

system. The first step of MaSE methodology is the

identification of the goal hierarchy diagram of the

OBSERVATIONANDSUPPORTINACOLLABORATIVEPEDAGOGICALDEVICE-AMulti-agentSystemto

ControlandMonitorPedagogicalActivities

137

Figure 2: Goal hierarchy diagram.

observation of use and of the support system. This

diagram is shown in figure 2.

3.2 The Agents of the System

The figure 3 presents an overview of the device,

with the different agents of the system. We

distinguish two types of user: the student and the

teacher, and three workspaces: the teacher’s space,

the student’s space and the group’s space. Every

user is associated to an agent that is located on the

server. This agent migrates on the user’s workstation

as soon as he/she connects. The agent is coded as a

Java program; applets are programs living on the

server and which run on the client. This technology

allows a user to run his/her agent directly from

his/her client. The superintendent of the space

groups lives on the server. This agent is active as

soon as one student of the group is connected. It

provides meta-information on the activities

(beginning date, end date, concerned persons, used

tools, etc.) and on the forums (beginning date, end

date, etc.). An agent supervises the interaction. It

oversees every communication tool (email, forum,

chat, blog) and supervises all the actions done during

the session. Every event is dated and commented.

We store information of each user: connections

into the various spaces, the activities done and

communications. By aggregating this information,

we can obtain (from an agent which manages the

group’s space) information that helps us to

appreciate the life of the group, the productivity of

the members and the level of realization of the

educational activities. The analysis of this

information, by an evaluation agent, allows to

estimate the lifecycle of the formation. This agent

can make objective decisions about modifications of

the calendar of activities.

The main agents of the system are:

• a-LEARN: It supervises the student’s space. It

allows the supervision of all activities of a student

and provides an overall evaluation of his behaviour

during a training session;

• a-TEACH: It supervises the teacher’s space. It

allows supervision of educational resources loading

in his/her logbook, of access to group’s logbooks

and of used tools to communicate with students.

• a-GROUP: It supervises the group’s space

during a session. It supervises actors' activities

during a session. It indicates the degree of respect,

the success rate, the start date and end date of an

activity. This agent provides the list of present

students and absent students in a group and must

provides statistics concerning the progression of

each activity. It reminds students about deadlines

and notifies the late groups by sending alerts.

• a-TOOL: It supervises the use of tools and

provides statistics on the use of space tools (Email,

Chat, Forum, blog, CVS, etc.).

• a-EVAL: This agent's role is to aggregate the

information collected, to structure them in order to

present it to the Evaluation module of

MACADDAM studio.

3.3 A Case Study

Let’s take the example of the G1 group. G1 has two

students Patrick and Sam who must conduct an

activity assigned by their teacher Michael.

This activity requires a maximum period of

completion of 3 days and consists of two tasks T1

and T2 which can be performed in parallel. The

teacher has uploaded the resources R1, R2 and R3 in

his logbook. These resources may assist students in

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

138

Figure 3: Main agents in their respective spaces.

carrying out the activity.

Table1: Chronology of actions on Michael’s space.

Chronology a-TEACH Mickeal’s space

J - 9h00 Patrick accesses the logbook of the teacher

J - 9h01 Patrick reading the activity to do

J - 9h15 Patrick Downloads R1

J - 9h17 Patrick Downloads R2

J - 9h24 Patrick Downloads R3

J+1 - 18h00 Patrick accesses the logbook of the teacher

J+1 - 18h01 Patrick uploads deliverables for the activity A

(task T1)

J+2 - 18h00 Patrick accesses the logbook of the teacher

J+2 - 18h01 Patrick uploads deliverables for the activity A

(task T2)

To begin the implementation of this activity,

Patrick connects on the logbook of the teacher and

downloads the available resources. He then accesses

the chat to send a message to Sam. He proposes to

be the leader of the group and offers a division of the

activity between them: he should make T1 and Sam

should make T2. While waiting for Sam’s answer,

he begins reading the downloaded documents and he

works on the completion of T1. At the end of the

next day, T1 is completed. Patrick uploads the

deliverables on the logbook of his group. He finally

receives an acceptance message from Sam. Patrick

offers to help him in the realization of T2 as he has

already finished the task T1. But despite this help

done via the chat, at the end of the 3 days, G1 has

made only 70% of T2. Part of the deliverables has

been uploaded on the logbook of the group.

Table 2: Chronology of actions on Patrick’s space and

Sam’s space.

Chronology a-LEARN Patrick’s space a-LEARN Sam’s

space

J - 9h50 Chat access

J - 9h54 Sends a message to Sam

J - 10h00 Reading R1 resource

J - 10h50 Reading R2 resource

J - 11h15 Reading R3 resource

J - 11h30 Consulting chat

J+1 - 17h00 Chat access

J+1 - 17h01 Read message

J+1 - 17h05 Sends a message to

Patrick

J+1 - 18h06 Consulting chat

J+1 - 18h10 Sends a message to Sam

J+2 -

8h to 17h

Discussion with Sam Discussion with

Patrick

J+2 - 17h45 Sends deliverable

Without the presence of the observation system

of use and assistance, the teacher would have

considered Patrick and Sam in the same way.

On the other hand, the supervisor agents of

Patrick’s space and Sam’s space were able to collect

the traces represented table 2.

The supervisor agent (a-GROUP) noted that

Patrick was present throughout the performance of

the activity in contrast to Sam who was absent for

about 2/3 of the time allocated to carry out the

activity.

Thus, the teacher is able to assess the amount of

work done by each student, their state (present,

absent, inactive), their degree of sociability, the

OBSERVATIONANDSUPPORTINACOLLABORATIVEPEDAGOGICALDEVICE-AMulti-agentSystemto

ControlandMonitorPedagogicalActivities

139

achievement level of the activity, etc. Alerts are sent

when something abnormal occurs.

4 ACCOMPLISHMENTS AND

FUTURE WORK

We have chosen to deploy the system on the Madkit

multi-agents platform because it is intended for the

development and the execution of multi-agents

systems and more particularly those based on

organizational criteria.

The MaSE methodology and the deployment of

communication mechanisms on the Madkit platform

allowed us to test the feasibility of the multi-agent

system for the observation of the learning system

and to validate our conceptual approach.

Several issues remain to be explored and

implemented. This includes the development of

mechanisms to make the device more autonomous

and proactive. Thus, the environment would be able

to prevent the teacher and the students when a group

has a bursting risk or is in a position of educational

failure. Collecting interaction traces of different

users does this.

Analysis of these traces is used to provide four

types of assessments:

• A group level assessment: status related to the

implementation of educational activities and to its

members’ status (present, absent, inactive). The

group supervisor agent that resides on the server

carries out this assessment.

• A student level assessment: assessing his/her

productivity related to the realization of the

pedagogical activities, his/her sociability that

indicates his/her communication level with other

members of the group. An inactive user connects to

the platform only to read messages.

• An activity level assessment: indicating the level

of achievement of an activity by the group.

• A teacher level assessment: evaluating the

predictive pedagogical scenario compared to the

effective scenario. This agent will interact with the

evaluation module of the MACADDAM studio.

REFERENCES

Al-Sakran, H. and Serguievskaia, I., 2008. Framework

Architecture of e-Loan Negotiation System, The IEEE

3rd International Conference ICTTA'08, Damascus,

Syria.

Caron, P.A., 2007. Web services plug-in to implement

"Dispositives" on Web 2.0 applications, in EC-

TEL’07, Second European Conference on Technology

Enhanced Learning, Springer LNCS Crete, Greece.

Downes, S, 2005. E-learning 2.0, eLearn Magazine, 17

October: http://www.elearnmag.org/subpage.cfm.

Ferber, J., 1999. Multi-Agent Systems. An Introduction to

Distributed Artificial Intelligence, Addison Wesley,

London.

Florea, A., Kayser, D., Pentiuc, S., El Fallah Segrounichi,

A., 2005. Agents Intelligents, http://turing.cs.pub.ro

/auf2/index.html.

Talon, B., Leclet, D., 2011. Towards a computer aided

pedagogical engineering, CSEDU’11, Netherlands.

Williams, B. J., Jacobs, J., 2004. Exploring the use of

blogs as learning spaces in the higher education sector,

Australasian Journal of Educational Technology

(AJET), Vol. 20, n°2, p 232-247.

Leclet, D., Talon, B., 2008. Assessment of a Method for

Designing E-Learning Devices, Proceedings of World

Conference on Educational Multimedia, Hypermedia

and Telecommunications, ED-MEDIA, Vienna,

Austria, AACE/ Springer-Verlag (Ed.), p 1-8.

Schneider, J. G., Johnston, L., Joyce, P., 2005. Curriculum

Development in educating Undergraduate Software

Engineers – Are Students being prepared for the

profession?, Australian Software Engineering

Conference, pp 314-323.

Settouti, L., Guin, N., Mille, A. and Luengo, V., 2010. A

Trace-Based Learner Modelling Framework for

Technology-Enhanced Learning Systems, 10

IEEE

International Conference on Advanced Learning

Technologies. Sousse, Tunisia. pp. 73-77.

Roux, J. P., 2004. Le travail en groupe à l'école, Les

cahiers pédagogiques, n°424: http://www.cahiers-

pedagogiques.com/IMG/pdf/Roux.pdf.

Zambonelli, F. Jennings, N. R., and Wooldridge, M.,

2003. Developing Multiagent Systems: The Gaia

Methodology, In ACM Transactions on Software

Engineering Methodology, 12(3):317-370, July.

Deloach, S., 2004. The MaSE Methodology, In

Methodologies and Software Engineering for Agent

Systems, Kluwer.

Tinio, V., 2002. ICT in education, World Summit on the

Information Society, http://en.wikibooks.org/wiki/

Category:ICT_in_Education.

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

140