Développement d'une plateforme d'apprentissage adaptatif basée sur l'IA pour l'éducation en ligne

Thesis Abstract

The rapid proliferation of online education necessitates innovative approaches to enhance learner engagement and personalized learning experiences, especially in the context of diverse learner needs and varying levels of prior knowledge. This study addresses the challenge of developing scalable, intelligent solutions that adapt instructional content dynamically to individual learner profiles, fostering improved educational outcomes. The primary aim is to design and validate a novel adaptive learning platform leveraging artificial intelligence (AI) techniques to tailor instruction in real-time, thereby bridging gaps in traditional e-learning environments. Specific objectives include analyzing pedagogical requirements for adaptive systems, developing an AI-driven recommendation engine, implementing a user-centered platform prototype, and evaluating its effectiveness through empirical testing. Employing a mixed-method research design, the study combines quantitative and qualitative data collection and analysis procedures. The target population comprises 300 adult learners enrolled in online degree programs at a regional university, selected through stratified random sampling to ensure representation across different disciplines and experience levels. Data collection instruments include standardized assessments of learner engagement and performance metrics, system usage logs, and semi-structured interviews with both learners and educators. The platform's recommendation algorithms are based on machine learning models, specifically collaborative filtering and reinforcement learning, integrated within a web-based interface. Validity and reliability of the instruments are established through pilot testing and Cronbach’s alpha analysis, respectively. Quantitative data are analyzed using multiple regression analysis, ANOVA, and system performance metrics, while qualitative data undergo thematic analysis to explore user perceptions and pedagogical implications. Expected findings indicate that the AI-powered adaptive platform significantly improves learner engagement, motivation, and academic performance compared to traditional static online courses. The platform demonstrates high predictive accuracy in content recommendations, with user satisfaction ratings exceeding 85%, and facilitates personalized learning pathways that accommodate diverse cognitive and motivational profiles. The results suggest that integrating AI techniques such as collaborative filtering and reinforcement learning effectively enhances the adaptability and scalability of online education systems. This research contributes to the field of educational technology by providing a comprehensive framework for developing intelligent, learner-centered online platforms grounded in pedagogical theory and AI. It advances existing knowledge on adaptive learning systems by empirically validating the effectiveness of hybrid machine learning models in real-world educational settings. The study also offers a practical prototype that can be adopted or further refined by educational institutions seeking to implement personalized learning environments at scale. The main conclusion emphasizes that AI-driven adaptive platforms hold substantial promise for transforming online education, fostering inclusivity and improved learning outcomes across diverse learner populations. Based on these findings, recommendations include integrating such platforms into mainstream curricula, investing in training educators to leverage adaptive technologies, and conducting longitudinal studies to assess long-term impacts on learner success. Future research directions proposed involve exploring the integration of natural language processing for real-time feedback and expanding the platform’s capabilities to support collaborative and social learning features. Overall, this study underscores the vital role of AI in shaping a more responsive, effective, and equitable online education landscape.