Applications of Models Learned from Data in Instructional Contexts
PRESENTER:
Kasia Muldner
Carleton University, Institute for Cognitive Science
ABSTRACT:
Tutoring systems are educational technologies that support students
through personalized instruction. To this end, these technologies rely on
models that recognize student states relevant to instructional activities,
such as creativity, affect, and effort. In this talk, I will describe our
work on building models from data collected in studies involving students
engaged in various instructional activities. I will focus on a series of
analyses involving the automatic extraction of linguistic features from
collaborative problem-solving activities to build models that predict
student creativity. I will also touch on work modelling other student
characteristics, including affect and effort-related outcomes.