|
Università degli studi di Udine
Dipartimento di Matematica e Informatica
|
Artificial Intelligence
Laboratory
Student Modeling Group
Goals
Student modeling represents one of the most important features which an
intelligent tutoring system needs to provide in order to adapt its
behavior to the specific traits of a student. In our work we have studied
knowledge-based techniques for building and maintaining the student model,
a structure describing the behavior, the knowledge, and the reasoning of
the student in a specific subject domain. The student modeling process
encompasses two main activities: analysis and interpretation of the
student's behavior, and management of the student model. Both tasks have
been investigated in our research work by proposing and integrating
different kinds of techniques:
Backward model tracing: analysis of the student's reasoning process by
reconstructing, step by step and in reverse order, the chain of reasoning
(s)he has followed in giving his/her answers.
Bug construction based on explanation based learning: dynamic generation of
possible student's incorrect domain concepts by using explanation-based
generalization techniques.
Truth maintenance techniques for managing the student model: exploitation
of an assumption-based truth maintenance mechanism in order to manage
alternative hypotheses concerning student's knowledge and to handle
possible contradictions observable in the student's behavior.
Temporal reasoning: modeling of changes in the student's knowledge by means
of an explicit representation of the temporal knowledge related to
student's beliefs.
Results
Many ideas and techniques concerning student modeling have been
experimented in ET, a prototype of intelligent tutoring system used to help
Italian students to master the core and the fine nuances of English verbs
usage. ET is the result of almost ten years of research work and relies on
a deep model of the knowledge underlying tense selection and verb
conjugation. ET proposes adequate exercises, analyzes the students
responses, hypothesizes the reasons underlying their mistakes, and makes
them aware of their misconceptions. A prototype version of the ET system
has been fully implemented in Prolog and runs on Macintosh computers (it is
available on floppy disk).
Main Publications
- Fum, D., Giangrandi, P., e Tasso, C. (1988) ET: an Intelligent Tutor for Foreign Language Teaching. Proceedings of the International Conference on Intelligent Tutoring Systems, Montreal, Canada.
- Fum, D., Giangrandi, P., e Tasso, C. (1988) Student Modeling Techniques in Foreign Language Tutoring. Proceedings of ECAI-88.
- Fum, D., Giangrandi, P., e Tasso, C. (1990) Backward Model Tracing: An
Explanation-Based Approach for Reconstructing Student Reasoning.
Proceedings of AAAI-90, Boston.
- Tasso, C., Fum, D., e Giangrandi, P. (1992) The Use of Explanation-Based Learning for Modeling Student Behavior in Foreign Language Tutoring. In Swartz, M.L, e Yazdani, M. (eds.) The bridge to International Communication: Intelligent Tutoring Systems for Second Language Learning. Springer-Verlag.
- Giangrandi, P. and Tasso, C. (1995) Truth Maintenance Techniques for
Modelling Student's Behaviour. Journal of AI and Education vol. 6 (2/3).
- Giangrandi, P. and Tasso, C. (1996) Modelling the temporal evolution of
student's knowledge. Proceedings of the European Conference of Artificial
Intelligence and Education.
Principal Investigators
Prototype
- the ET project in various version
Events
The group will present some of the latest results concerning Temporal Student Modeling at UM97 - User Modeling '97 and AI&ED97- International Conference on Artificial Intelligence in Education, Kobe, Japan, August 1997.
Research Partnership
Several research activities on intelligent tutoring systems have been
carried out in collaboration with Prof. Danilo FUM (Department of
Psychology - University of Trieste, Italy).
Previous page
AI Lab
Goals