Learning science knows more than physical environments can currently use. Public robotics has been deployed for two decades without transferable infrastructure. P.A.I.R. Lab is working on both gaps.
Learning science has established clearly that cognitive state — where a learner actually is in their thinking at a given moment — determines the quality of what they take away from an experience more than the content they encounter. Adaptive and intelligent tutoring systems have pursued this insight for decades, with meaningful results in structured digital environments. The open problem is the physical world.
That is the problem Project Octopus is working on. Not adaptive learning as a category — that literature is deep and we build on it — but the specific infrastructure gap between what learning science knows and what can currently be deployed in a physical experiential environment at scale, without a skilled human mediator at every point of contact.
The institutional deployment problem in public robotics is well documented and consistently unsolved. Social robots have been deployed in public experience environments for over two decades — the Smithsonian, CosmoCaixa in Barcelona, science museums across Japan. Each deployment required a dedicated engineering team, a proprietary software stack locked to a single manufacturer, and sustained institutional investment to keep running. None of that work transferred between institutions.
Project ROSE is building that layer. An open, modular operating system for deploying robots in public experience environments.