In the first year, project teams focus on a basic research question connecting some elements of human learning, data-science, machine learning, human-machine interaction, or education. The goal is to learn cross-disciplinary concepts and communication while completing publishable work in the first year.
In this example students from engineering and educational psychology collaborate to understand how undergraduate chemistry students perceive diagrams of molecular structures. Students judge perceptual similarities in (1) and adaptive algorithms efficiently compute the latent structure governing judgments, shown in (2). Sparse optimization methods are applied to learn which diagram features in (3) are most important for generating the perception of similarity. Comparing features used by experts and novices can illuminate how these groups differ and provide targets for instructional interventions from teachers and intelligent tutoring systems.