Designing Personalized Learning Environments throughMonitoring and Guiding User Interactions with Code andNatural Language

Learning the vocabulary of a new language and a new programming API are similar in multiple ways. In this paper we evaluate several of the similarities and show that based on them we can design systems that can guide the learner towards improving their knowledge without an external tutor or preset curriculum. Instead, the class of systems we propose here are based on automated approaches of building maps of knowledge of the domain by mining repositories. By intersecting this knowledge with models of learner knowledge built by observing past learner interactions with artifacts of the domain we can generate highly personalized learning guidance.