Communication Skills Platform for LeArning via Role-playing games (CoSPLAyR)

The Communication Skills Platform for LeArning via Role-playing games (CoSPLAyR) is a set of software tools and embeddable modules that, together, provide a platform for developing intelligent virtual tutoring apps interpersonal communication skills.

CoSPLAyR consists of:

  • A curriculum-driven authoring tool that combines science-of-learning models of curriculum with narrative-models of experiential practice design. It guides entry of curriculum and instructional design elements (objectives, performance measures, didactic instruction content, tutoring scaffolding policy and content) and of domain descriptions (of practice elements and scenario, narrative maps of different paths through a practice scenario, descriptions of characters the learn will interact with). The authoring tool then guides the final step of to creating multi-path practice interactions mapped to the various points where learning objectives come into play. The final product of the tool is a data representation (in JSON) of all elements of the curriculum and practice interaction descriptions.
  • A tutoring management controller that guides an app or website that steps a learner through the learning curriculum, providing access to didactic instruction and practice interactions appropriate for each objective or level of the curriculum. The controller also allows the learner to select (or repeat) practice interactions, presents the learners with their progress in demonstrating skills in each object, and gives them access to more advanced objectives/levels once a given skill threshold of performance is attained. The controller also dispenses reinforcing rewards (badges, commendations) based on skill performance. [The simulation of the interactions is provided externally, e.g. through a game engine or interconnected lattice of video interaction segments, all from the learner’s point of view.]
  • An intelligent tutoring and learner-modeling module that infers the likelihood that the current user/learner has attained competence and/or mastery of each individual objective, based on an Bayesian inference model using the sequence of choices made relative to a given objective at points in the exercise where that objective is relevant. The inferred learning state is used to make available proactive coaching relative to each relevant objected at every choice point, and to make available responsive feedback relative to each relevant objected after every choice is made. The skill estimates from the learner model are used by the tutoring management controller to determine when the learn can advance to a next level, needs remedial instruction relative to a specific objective, and to award rewards.

The simulation of the interactions in each practice interaction needs to be developed separately for each new application of CoSPLAyR, although the Authoring tool provide very structured specifications for developing that simulation.