Adaptive Skill Transfer: Reinforcement-Learned Scaffolding for Cognitive Personalization in Adult Learning

As automation and AI reshape work, adults must acquire new skills that demand abstract and data-driven reasoning. Current learning systems often ignore the general cognitive abilities adults already possess—such as reasoning, pattern recognition, and strategic planning. This project aims to model how these abilities can accelerate domain learning and how adaptive AI systems can make this transfer explicit and efficient. Implemented within Aacharya, an adaptive learning platform, the research combines cognitive assessment, modeling, and personalized tutoring to adapt not only content but cognition itself, transforming raw materials into interactive lessons tailored to each learner’s thinking style.​

A key innovation is embedding simple interactive games into Aacharya to assess reasoning, planning, and reflection before instruction begins. These games generate cognitive profiles that reveal how learners approach problems, which then guide personalized scaffolds linking their natural reasoning patterns to new domain concepts. The project hypothesizes that leveraging these profiles will significantly improve learning speed and retention. Through integrated components—cognitive modeling, adaptive scaffolding, and transfer efficiency evaluation—the research will build a computational theory of adaptive skill transfer. Beyond academic contributions, it addresses a critical societal need: enabling adults to learn faster by turning existing cognitive strengths into actionable pathways for reskilling.​

​Research Impact: ​Transforming adult reskilling from simple content delivery into a personalized, AI-driven bridge that leverages existing reasoning strengths to accelerate the mastery of complex skills