ECHO Project (e-Learning)
E.C.H.O. = Étude Comportementale des Hommilières pour l'Optimisation (Behaviourist Study of Man-Hill for Optimization)
Collaboration : LIL Laboratory, Calais, France and Paraschool Compagny, Paris, France
Goal :
The aim was to improve the e-learning software of Paraschool Company by using evolutionary computation technics :
Apply a variation of the Ant Colony Optimization algorithms to increase the learning and
Create an adaptative interaction with the user to rise her/his interest.
Description :
ACO (developped after the observation of ant-hills) uses virtual ants to find minimal path in a graph. In the Paraschool system, the very large number of students triggered the idea to apply a similar technique using real students rather than virtual ants, with the aim of optimising pedagogical paths traversing a set of educational topics.
Real-size experimentations have shown that ant-hill optimization techniques developped in Paraschool do not directly apply because students do not behave like artificial ants. The concept of an artificial "student-hill", or more generally "man-hill" has been introduced.
New concepts have then been introduced to overcome differences of behaviours and match the pedagogical goals of Paraschool :
The "Erosion" concept of pheromones to overcome the sporadic activity of students
The design of the fitness function by mixing global goals and personnal wishes
The use of Elo Notation to propose strength-adapted drills
The evolution of an avatar to increase the interaction with the user
The analysis shows that the "man-hill" paradigm is stable over-time and behaves as expected. Therefore the advantages are numerous :
The learning of students is improved by the navigation among exercices.
The teachers have a new audit tool that detect semantic errors over one thousands drills and propose new pedagogical paths.
Fitness function :
Arcs interconnecting items bear values that are used to compute a fitness function that evaluates how nice it would be for a particular student to follow a particular arc. The fitness function group together
Pedagogical Data: The pedagogical weight represents a static advice given by the teacher.
Behavior Data (Pheromones) : There are two types of pheromones representing success and failure. These data bring to the system an estimation of the success rate and of the number of visits on the differents pedagogical paths.
Individual Data : Those parameters allow to personnalize the suggested paths to each student.
This figure sums up the concepts used by the man-hill paradigm.
Navigation :
This figure is a snapshot displaying the result of the student and the suggestions found by the man-hill paradigm.
Publications :
[2006] Étude Comportementale des Hommilières pour l'Optimisation Grégory Valigiani, Évelyne Lutton, Pierre Collet In EpiNet, EPI Electronic Magazine, n' 83, march 2006 Refereed Journal Article
[2006] Automatic Rating Process to Audit a Man-Hill Grégory Valigiani, Évelyne Lutton, Yannick Jamont, Raphaël Biojout and Pierre Collet In WSEAS Transactions on Advances in Engineering Education, p1-7, Issue 1, Volume 3, January, ISSN 1790-1979, 2006. Refereed Journal Article
[2005] Evaluating a Real-Size Man-Hill Grégory Valigiani, Évelyne Lutton, Yannick Jamont, Raphaël Biojout, Pierre Collet In ECE'WSEAS'05E, 17-19 November, Miami, Florida, USA, 2005. Refereed Conference Contribution
[2005] Experimenting with a Real-Size Man-Hill to Optimize Pedagogical Paths Grégory Valigiani, Raphaël Biojout, Yannick Jamont, Évelyne Lutton, Pierre Collet In SAC'05 Proceedings, 13-17 mars, Santa Fe, New Mexico, USA, 2005. Refereed Conference Contribution
[2003] Application de l'optimisation par colonies de fourmis à la structuration automatique de parcours pédagogiques Pierre Collet and Yann Semet In Revue Enseignement Public et Informatique, Septembre, 2003. Refereed Journal Article
[2003] Ant Colony Optimisation for E-Learning: Observing the Emergence of Pedagogic Suggestions Pierre Collet, Evelyne Lutton and Yann Semet In IEEE Swarm Intelligence Symposium, Indianapolis, USA, April 24-26., 2003. Refereed Conference Contribution
[2003] Artificial Ant Colonies and E-Learning: An Optimisation of Pedagogical Paths Yann Semet, Evelyne Lutton, Pierre Collet, Yannick Jamont and Raphael Biojout In HCII'03 - Human Computer Interaction International, 2003. Refereed Conference Contribution