A mixed-initiative intelligent tutoring system: Based on learning from demonstration
Conference Paper
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
We present the design and evaluation of the framework of a Mixed-Initiative Intelligent Tutoring System that augments existing tutoring systems by integrating two interactive modes: instructor-student, and intelligent tutor-student. These interactive modes are intended to support students in well- and ill-defined problem solving. In this paper we discuss the use of the Learning from Demonstration approach to derive the solution paths and the appropriate tutorial actions in response to observed student behavior and instructor intervention in the cybersecurity domain. Our method aims to discover large portions of domain and tutoring knowledge from instructors' interactions with students at run time. We describe the use of a Weighted Markov Model approach for data representation for sequential data. Our experimental results indicate that the proposed technique is useful for data sets of sequences.