Paper: Simple but effective feedback generation to tutor abstract problem solving

ACL ID W08-1114
Title Simple but effective feedback generation to tutor abstract problem solving
Venue International Conference on Natural Language Generation
Session Main Conference
Year 2008
Authors

problem solving Xin Lu, Barbara Di Eugenio, Stellan Ohlsson and Davide Fossati University of Illinois at Chicago Chicago, IL 60607, USA xinlu@northsideinc.com bdieugen,stellan,dfossa1@uic.edu Abstract To generate natural language feedback for an intelligent tutoring system, we developed a simple planning model with a distinguishing feature: its plan operators are derived auto- matically, on the basis of the association rules mined from our tutorial dialog corpus. Auto- matically mined rules are also used for real- ization. We evaluated 5 different versions of a system that tutors on an abstract sequence learning task. The version that uses our plan- ning framework is significantly more effective thantheotherfourversions. Wecomparedthis version to the human tutors we employed in our tutoria...