This chapter presents how Machine Learning Techniques can effectively contribute to improve the quality of interactions in Guided Discovery Tutoring Environments (GDTE) . We review several approaches to integrate Machine Learning in ITS. Most of these approaches use concept learning from examples to maintain a Student Model. We go along presenting an alternative use of induction techniques to learn concepts on the same data that are presented to the learner. We present on a concrete example how this approach is integrated in a GDTE called MEMOCAR, a Computer Aided Language Learning System for Chinese characters. Three main types of activity are identified in MEMOCAR: familiarization with Chinese characters, collaborative discovery of similarities between characters and exercises to test characters acquisition. The stage of familiarization is supported by exploration of hyperdata whilst collaborative discovery and exercises' diagnosis are supported by a tool based on CHARADE, a top-down induction system. Such integration offers a new alternative to the complex problem of making Guided Discovery Tutoring Environment more collaborative. Resume: Le theme aborde dans ce rapport est celui de l'utilisation de techniques d’apprentissage symbolique automatique (ASA) pour l'amelioration des interactions dans les environnements d’enseignement par la decouverte assiste par ordinateur. Nous presentons differents environnements integrant des techniques d'apprentissage symbolique automatique; principalement pour construire et/ou mettre a jour un modele de l'eleve. Nous presentons ensuite une integration originale de l’ASA dans le cadre d'un environnement dedie a l'apprentissage des caracteres chinois par la decouverte: MEMOCAR. Notre approche consiste a utiliser l'ASA pour faire apprendre au systeme sur les memes donnees que l'eleve, et a utiliser les resultats de cet apprentissage pour construire des interactions. Deux des principales activites du systeme MEMOCAR, la decouverte par l'apprenant de similarites entre caracteres et les exercices de test de l’apprentissage sont ainsi basees sur l'adaptation d'un algorithme inductif: CHARADE. Ce type d’integration de l'ASA dans les environnements d'apprentissage par la decouverte s'insere dans les recherches visant a rendre de tels environnements plus cooperatifs.
[1]
Mark Elsom-Cook.
Guided discovery tutoring : a framework for ICAI research
,
1990
.
[2]
Steven Pinker,et al.
Language acquisition
,
2003
.
[3]
R. Michalski,et al.
Learning from Observation: Conceptual Clustering
,
1983
.
[4]
Tom M. Mitchell,et al.
Generalization as Search
,
2002
.
[5]
Stellan Ohlsson,et al.
Automated Cognitive Modeling
,
1984,
AAAI.
[6]
Allan Collins,et al.
The goal structure of a socratic tutor
,
1977,
ACM Annual Conference.
[7]
Ryszard S. Michalski,et al.
Toward a Unified Theory of Learning: An Outline of Basic Ideas
,
1991
.
[8]
Douglas L. Medin,et al.
A two-stage model of category construction ☆
,
1992
.
[9]
Michelle Joab,et al.
Diagnostic cognitif de l'apprenant par apprentissage symbolique
,
1992,
Intelligent Tutoring Systems.
[10]
Kurt Van Lehn.
Two Pseudo-Students: Applications of Machine Learning to Formative Evaluation
,
1990
.
[11]
J. Ross Quinlan,et al.
Induction of Decision Trees
,
1986,
Machine Learning.
[12]
Tom M. Mitchell,et al.
The Need for Biases in Learning Generalizations
,
2007
.
[13]
Kurt VanLehn,et al.
Applications of simulated students: an exploration
,
1994
.
[14]
Jean-Gabriel Ganascia.
CHARADE: A Rule System Learning System
,
1987,
IJCAI.
[15]
J. G. Ganascia,et al.
Deriving the learning bias from rule properties
,
1991
.
[16]
Woo-Kyoung Ahn,et al.
A Two-Stage Model of Category Construction
,
1992,
Cogn. Sci..
[17]
Michael Grüninger,et al.
Introduction
,
2002,
CACM.
[18]
Michel Caillot,et al.
Student modelling by a knowledge‐based system
,
1986,
Comput. Intell..
[19]
Derek H. Sleeman.
Inferring (Mal) Rules from Pupil's Protocols
,
1982,
ECAI.
[20]
Mitsuru Ikeda,et al.
THEMIS: a nonmonotonic inductive student modeling system
,
1994
.
[21]
John F. Sowa,et al.
Conceptual Structures: Information Processing in Mind and Machine
,
1983
.
[22]
Jean-Gabriel Ganascia,et al.
Selective Reformulation of Examples in Concept Learning
,
1994,
ICML.