A Machine Learning Approach to Personalizing Learning by Reading Experiences Based on Individual Cognitive Profiles and Preferences

Abstract

Modern educational contexts demand adaptive methods that cater to individual learning styles, cognitive capacities, and preferences. This paper addresses the challenge of personalizing reading experiences for diverse learners by leveraging a machine learning architecture grounded in cognitive psychology and advanced user modeling. The proposed approach systematically captures key cognitive traits, including working memory capacity, information-processing speed, and domain-specific prior knowledge, to tailor reading materials and strategies to each learner’s profile. By mapping user interactions with various textual complexities, semantic structures, and contextual cues, the system dynamically adjusts presentation, pacing, and conceptual scaffolding. A rich set of features, encompassing linguistic and psychometric indicators, contributes to building robust user representations. In contrast to purely content-based adaptation, our framework explicitly integrates a cognitive layer that mediates user intentions, engagement levels, and meta-cognitive strategies for comprehension monitoring. A combination of probabilistic modeling and deep learning is employed to infer latent cognitive states, predict performance trajectories, and fine-tune content difficulty. Preliminary evaluations show promising improvements in reading comprehension, engagement, and sustained learning gains, while also revealing the importance of personalization that aligns with fine-grained cognitive profiles. This paper offers a new perspective for adaptive learning systems, guiding future developments toward more inclusive, evidence-based, and cognitively aligned educational technologies.