Towards Intelligent Tutoring with Mathematical Sketching

Опубликовано 17 августа 2016, 2:02

Mathematical sketching lets users create dynamic illustrations through the association of hand written mathematics and free-form drawings. These dynamic illustrations have the potential to provide useful insights into a variety of abstract concepts in Physics. Thus, mathematical sketching lends itself, with its pencil-and-paper-based interaction approach, to be a natural interface to intelligent tutoring applications. Although mathematical sketching supports a wide variety of dynamic illustrations, it is significantly limited because users must directly specify how objects behave with position and/or rotation functions of time. In contrast, problems students are asked to solve infrequently conform to this dynamic illustration creation scheme. Thus, mathematical sketching needs to be broadened to include inferencing capabilities to make a proper dynamic illustration, given information that is only indirectly or partially related to providing a behavioral specification for the animation. In this talk, I will present our current progress toward empowering mathematical sketching to support a more robust set of dynamic illustrations that would cover the majority of concepts found in an introductory Physics course.