Interoperability standards let learning applications share data and connect infrastructure, thereby reducing development costs, administrative burden, and user friction. But users can still face significant burden from having to move between applications, and from having to figure out how to take advantage of potential synergies. We propose extending interoperability standards to facilitate shared user interfaces and session management services. This will let applications leverage differences in how and when they are used to achieve a more unified user experience (UX) and better learning outcomes than is possible with only native capabilities. To demonstrate this concept, we integrated two complementary applications. PERLS is a mobile Personal Assistant for Learning that recommends learning activities across a range of applications and devices, and provides a coherent user experience as users progress through learning trajectories varying in topic, timespan, intensity, and (in)formality. PALMS is a web service that uses flashcard-like interactions to train visual knowledge. It accelerates learning by adaptively optimizing the content sequence based on a model of learning rate and retention. We demonstrated UX unification by allowing PERLS to recommend PALMs content, launch it, and then hand off interaction control in a way that left users unaware of switching applications. We demonstrated learning advantage by allowing PALMs to schedule future interactions within PERLS, effectively spacing practice out over a prolonged time scale in order to improve retention. The talk will review this, as well as more recent work with ADL’s Total Learning Architecture, in which PERLS recommended content and managed workflows between 7 learning applications and 3 devices.
Michael Freed, Ph.D. is a Program Director with SRI International in the Artificial Intelligence Center. His research focuses on augmenting human capabilities for learning, health care, and work productivity using intelligent personal assistants and other human centered computing technologies.