Adaptive support for coaching meta-cognitive skills

We describe a computational framework designed to provide adaptive support to learning from examples by coaching the meta-cognitive skill known as self-explanation - generating explanations to oneself to clarify an example solution. The framework includes an interface to scaffold self-explanation, a probabilistic student model and a coaching component. The probabilistic student model allows the coach to change the interface scaffolding in order to improve the student's studying behavior. We discuss how the types of tailoring and scaffolding that the framework enables can be extended to generate more adaptive support to effectively learning from examples.