UC Santa Cruz

The idea that interactive computational media could be used to teach formal knowledge through concretization was articulated by Seymour Papert. The design of learning systems could similarly employ implicit learning\textemdash by way of immersive environments\textemdash for improved acquisition, retention, and transfer of more complex forms of knowledge and expertise. Combining such environments with intelligent tutoring systems (ITS) holds potential for synergies that might accelerate learning beyond what is possible individually. Furthermore, guided-inquiry and active-learning pedagogical approaches are well suited to form the nexus between these software domains. The integration of these and other sub-domains represents a new domain I call ``LIVE-Tutor''. Conceived through an affordance-based design approach and informed by embodied interaction ideals, this domain offers a tutorial model for immersive hybrid education. In particular, it promotes implementation of adaptive pedagogical methods to provide social support for tutor-to-student and peer-to-peer instruction in inquiry-based activities (such as science labs). This thesis communicates the basic domain definition, the process taken in my formulation of a guiding framework for it, and describes the creation and evaluation a prototype application which was used to validate and refine the framework. A new construct in HCI, Domain Usefulness Experience (DUX), is introduced as a theoretical basis for the framework, and a novel ITS approach to teaching interventions is also presented; one which leverages behavioral data of domain experts and learners as inputs to the construction of a reward function applied within a constraint based, model-free, on-line reinforcement learning algorithm I call ``IntelliHints''.