The Future of Education

Technology is espoused as a tool that can improve access to educational resources and allow for learning in more personalized and compelling ways. Additionally, it is largely recognized that the gap in science learning is widening and that interventions for increased representation by all learners is critical to the future development of knowledge and the economy. It is imperative that young learners are engaged in science learning to ensure increased interest in science, technology, engineering, and mathematics fields. All too often communities in poverty do not have exposure to resources that allow for personalized learning environments and supports for all types of learners. This presentations offers participants a new way of thinking about the use of games as a cognitive diagnostic platform for educators to respond to individual learners. Specifically, an inclusive adaptation approach for a web-based physics learning game (“RoboBall”) currently being developed for first, second, and third graders is discussed.

In this concept paper considering the effective implementation of game adaptation for a (Science, Technology, Engineering, Mathematics) STEM educational environment, the need for a multi-pronged approach to adaptation is presented. To that end, the recommend framework is an evolving, iterative, research-based approach that uses both player (learning type) profiling and experience profiling for content generation. A semantic generation framework model has been identified to enable game adaptation. A strategic plan for adaptation includes levels of sophistication to “ramp-up” from a low resolution to a high-resolution player experience. Earlier stages would require up front (in game) diagnostics to enable player profiling while the system requires human agent supervision to train it (even as practices are tested out in classroom settings). With continuous refinement and learning of the system made possible through crowd-sourced quantities of data, a high level of automated standalone could be possible. With this framework, play-testing a game would allow for the ability to use content utility models to create semantic gameplay descriptions. Much of the game performance measures will rely on User-Initiated Events (UIE) but may be coupled with data derived from pre-game diagnostics and backend metrics tracking to contextualize a specific learner’s profile.

This paper presents an overview of our purpose and approach to adaptation, offers considerations of current trends in adaptation approaches in serious games, describes examples of individual user adaptation per the differentiated learning profiles approach, discusses assessment in brief, and conclude with considerations of the potential contribution of this work.