Cognitive-Inspired Adaptive Learning Models for Personalized Digital Education

Abstract

Individualised digital learning platforms are subjecting more and more data-driven adaptive learning methods to support the learner diversity in knowledge, learning speed and interaction. Most of the current adaptive strategies, however, are performed largely on a performance-optimization basis, having little foundation on well-currently held cognitive and neural theories of human learning. This weakness makes them less interpretable and might limit their ability to learn in the long term. The project is a cognitive-inspired adaptive learning framework, which in its turn suggests incorporating into the personalised learning models some fundamental concepts of the cognitive science, i.e. working memory limitations, attention control, cognitive load use, and reinforcement-based feedback. The model based approach of the research involves mapping of cognitive constructs to computational adaptation processes in order to allow cognitively realistic personalization over and above data-focused optimization. The suggested scheme includes a learner cognitive state model, a cognitively annotated content representation module, an adaptive decision engine as well as a feedback regulation component. The strategies of personalization are based on approximate cognitive states aimed at dynamically varying the order of content, teaching speed, and feedback timing. An evaluation plan based on simulation analysis, which is based on cognitive parameters, is used to determine the effectiveness of the learning, the stability of the engagement and the retention behaviour in comparison with the traditional performance-based adaptive systems. These findings suggest that cognitively informed adaptation lessens the cognitive load, levels off the engagement of the learner and enhances the retention of knowledge during longer periods of evaluation. The research finds that the principle that incorporates cognitive and neural principles into adaptive learning design will not only increase theoretical consistency but also educational performance. This literature has added a cognitively based ground on the building of the next-generation customised digital learning structures to the forum of cognitive and neural research.