by Andrea Bunt, Cristina Conati, Michael Huggett, Kasia Muldner
Open learning environments can be beneficial for learning in ways not available in more tutor-controlled systems, because of the active role the learner plays in knowledge acquisition. However, it has been shown that not all learners are proficient in unconstrained exploration, restricting their ability to learn effectively in these environments. In this talk we present the Adaptive Coach for Exploration (ACE), a prototype computational framework that supports active exploration in an open learning environment by providing tailored support to overcome specific student difficulties.
ACE provides students with a highly-graphical, exploratory learning environment in the domain of mathematical functions. A Student Model assesses student knowledge and exploratory behaviour using a Bayesian network; ACE's Coach uses this assessment to generate tailored hints that support the exploratory behaviour of those students who would otherwise have trouble learning in an unsupervised environment.
After describing ACE's components, we present the promising results of a preliminary user study that gauges the system's effectiveness.