Research




     My research is in building more effective and engaging educational systems. This includes modeling the student's cognitive and affective states, planning and decision-making in real-world environments, involving the student in game-playing and other immersive situations, and natural language dialog.

    I work as a Cognitive Scientist for Carnegie Learning developing Cognitive Tutors, currently in math and English domains.  I believe these tutors are best in class for delivering proven, effective instruction.  They are currently in use by over half a million students in thousands of schools. 

    I earned my PhD in Intelligent Systems at the University of Pittsburgh.  My advisor was Kurt VanLehn.  I worked at the Learning Research and Development Center, a multidisciplinary research center for understanding and improving learning.

    For my PhD work, I developed DT Tutor, a decision-theoretic approach for selecting tutorial discourse actions.  This approach, embodied in a dynamic decision network, ensures that DT Tutor's action selections are optimal given its beliefs and objectives.  DT Tutor considers a rich set of tutorial state attributes in a quest to approach some of the sensitivity and subtlety exhibited by human tutors.  These attributes include the discourse state, progress on the tutorial task (e.g., solving problems), and the student's knowledge, focus of attention, and affective state.

    DT Tutor can either stand alone with the addition of suitable user interface, or act as an action selection engine for an existing tutoring system.  For my PhD project, I developed the Calculus Related Rates Tutor, or Calculus Tutor for short, a decision-theoretic intelligent tutoring system (ITS) for teaching calculus related-rates problems.  My dissertation evaluated the Calculus Tutor in the first comparison of a decision-theoretic ITS with a non-trivial competitor. 

    Previously, I developed RTDT (Reading Tutor, Decision-Theoretic) to test the feasibility of adding the DT Tutor technology to an existing tutoring system, Project Listen's Reading Tutor, in an entirely different domain.  The Reading Tutor listens to children as they read aloud and provides coaching using speech and graphical user interface actions.  RTDT was not integrated with the Reading Tutor, but combining the systems appeared to be feasible.
Comments