Open Global Learning Infrastructure: Authors perspective
An adaptive Multi-Agent based Architecture for Engineering Education
1. An adaptive Multi-Agent based
Architecture for
Engineering Education
Dunia Inés Jara, Paola Sarango Lapo, Miguel Rodríguez Artacho
UTPL, Loja (Ecuador)
UNED University, Madrid (Spain)
http://www.utpl.edu.ec
2. An adaptive Multi-Agent based Architecture
for Engineering Education
∗ Introduction
∗ Adaptative logical architecture proposed for moodle
∗ Tutor Module
∗ Knowledge Base Module
∗ Student Module
∗ Interface Module for Users
∗ Future projects
∗ Conclusions
3. Introduction
• VLE – Moodle
▫ Open source with tendency to an Adaptative Educational System (AES)
▫ Moodle is based on three main components:
The professor, the classroom, the student
Classroom
Database
Guidelines
Interface Activities Interface
Resources
Teacher Student
• The proposed architecture is based on the main areas of adaptation
defined in Brusilovsky (1996), providing presentation and navigation
adaptation using intelligent agents associated to different modules in
Moodle.
4. Objectives
∗ Adaptive Navigational support
∗ Ex. Links
∗ The better next
∗ Link hidding
∗ Adaptive collaboration suport
∗ Group creation
∗ Automatic share of information
∗ Presentation support
∗ Prerrequisite of a given task
∗ Tool sorting (according to priorities)
5. Tutor Module
Supported by the instructional design
The tutor modeling agent has been designed to perform the
following functions:
•Didactical-Pedagogical. (teaching style)
•Tutor Modeling. (implementation of contents)
6. Knowledge Base Module
• Initial knowledge of the system, expressed in inference
rules or probability distributions, these are used by the
agent to infer a conclusion or new knowledge, used various
information sources.
aps taxo -Courses
p tual m n omie
conce s -Enrolled students
student’s perception -Virtual library
environme
personal
nt data -LO repository
data
Agents
information
interaction on actions
data instructional
design
the
sau ies
r us olog
ont
7. Knowledge base inference
Axiom Description Concepts Relationship Logic representation
A={x/x is an area}
C={y/y is a degree}
One area contains one
AC = {x/y y>=1}
Area/Degree of more degrees Area has
A={x/x is a subject}
C={y/y is a content}
A subject is defined by
AC={y/y defines a subject }
Subject one or more contents Subject Study
The set AC is a relationship between A and C and {y/y defines a subject} is “x is a subject and ‘y’ is a
content then “y” defines an “x”, that is to say “y” defines a subject
8. Student Module
The agent for student
modeling
performs some functions:
•Creation of Student
Models.
•User information
Update.
9. Intelligent Agent for Student Modeling
Multiagent System
Agent1: Monitoring Agent VLE Agent 2: Student Modeling Agent
Access to Resources and activities Collection the Interaction data
Task Algorithm
Interaction level in the course C4.5 (Decission tree
generation)
Interaction level with the Bayesian Net
resource
Interaction level in the VLE J48 www.cs.waikato.ac.nz/ml/weka/
10. ADAPTATIVE LOGICAL ARCHITECTURE
PROPOSED FOR MOODLE
-Competencies
- Assesment
- User (student)
tracking
- Institutional
approach
11. Interface Module for Users
It shows all the information to the students, trying to capture their
attention and keeping them motivated, through redaction of
messages type “Inverted Pyramids” .
The objective of this agent is to determine the best interface
to be offered to each user based on the hardware and
software used for the connection.
12. Intelligent Agent for making instructional
decisions (ToDei)
The objective of this intelligent agent is to fulfill these functions as well as to transmit
the content to the user.
Furthermore, considering the characteristics and greatest needs, it decides the best
way to offer information generated in this process
13. CONCLUSIONS
• Different Inference approaches for different domains
(tutor, student, navigation, interface,..)
• Use of the Moodle information model to track actors
activity
• Evaluation not developed, but tracking is persistent
• The ToDei agent constitutes the main component inside
this architecture since it allows visualization of the adaptive
effect generated by the interaction of the components.
• Work in progress. Main development is knowledge base
interface information model and inference system
14. An adaptive Multi-Agent based
Architecture for
Engineering Education
Dunia Inés Jara, Paola Sarango Lapo, Miguel Rodríguez Artacho
UTPL, Loja (Ecuador)
UNED University, Madrid (Spain)
http://www.utpl.edu.ec
Notas del editor
+
A la representación lógica se la lee así: "x" tal que "x" es una asignatura, es decir cualquier valor que x puede tomar de un conjunto de dominio, es una asignatura, "y" tal que "y" es un contenido, es decir cualquier valor que y puede tomar de un conjunto de dominio, es un contenido, El conjunto AC es la relación entre A y C y {y/y define una asignatura} quiere decir que si "x"es una asignatura y "y" es un contenido entonces "y" define a "x", o "y" define a una asignatura..
Falta un contexto del proyecto y un diagrama de contexto con Moodle. Habría que comentar qué objetivos tenemos antes de poner esta transparencia