Expert Tutoring and Natural Language Feedback in Intelligent Tutoring Systems

Location: SEO 1000
Time: 2:00 pm, Thursday, June 21, 2007.
Speaker: Xin Lu
Advisor: Barbara Di Eugenio
Committee: Barbara Di Eugenio, Bing Liu, Tom Moher, Stellan Ohlsson, Martha
Evens (IIT)

Abstract:

Intelligent tutoring systems can provide benefits of one-on-one instruction automatically and cost effectively. To make the intelligent tutoring systems as effective as expert human tutors, this research aims at investigating what type of natural language feedback an intelligent tutoring system should provide and how to implement the feedback generation to engender significantly more learning than unsupervised practice. This research demonstrates the utility of a computational model of expert tutoring in generating effective natural language feedback in intelligent tutoring systems.

This presentation will start from a comprehensive study of the difference between one expert tutor and two non-expert tutors in effectiveness, behavior and language. Then it will present a rule-based model of expert tutoring which takes advantage of a machine learning technique, Classification based on Associations. The tutorial rules are automatically learned from a set of annotated tutorial dialogues, to model how the expert tutor makes decisions on tutors attitude, domain concepts and problem
scopes to focus on, and tutor moves. This presentation will also describe a framework of feedback generation with 3-tier probabilistic planning to employ the model of expert tutoring in the natural language feedback generation for intelligent tutoring systems. The 3-tier planning automatically generates, selects and monitors plans for generating
effective tutorial feedback based on the rule-based model and the information state which keeps track of the interaction in the intelligent tutoring system. At last, an evaluation of the framework will be discussed.