There is evidence from cognitive science research that sketching can improve student learning, by providing a spatially-oriented means of expression. Sketching is underutilized in education, and especially in intelligent tutoring systems, because sketches are difficult to grade. In the Spatial Intelligence and Learning Center, we tackled this problem by developing CogSketch, a cognitive model of high-level human visual representation and problem-solving. This model was developed to provide part of the science base needed for using sketching broadly in education, and to ensure it was useful for this, CogSketch is also a platform for sketch-based educational software. While the two CogSketch-based educational systems we created went well beyond the state of the art for visual processing and the use of analogy in generating advice for tutoring, they did not exploit other techniques that are well-known in the intelligent tutoring community. We propose to explore the integration of sketching into ITS more broadly by integrating CogSketch and Companions into the GIFT cloud, and creating a tutor for Simple Machines (e.g. levers, pulleys, gears, hydraulics), a highly spatial domain that is relevant both to STEM education and to Army training.
|Effective start/end date||7/21/17 → 9/12/19|
- Army Research Office (W911NF-17-2-0054-P00002)