Développement d'une plateforme d'apprentissage adaptatif basée sur l'intelligence artificielle

Thesis Abstract

The rapid advancement of information and communication technologies has transformed educational practices, emphasizing the need for personalized and efficient learning environments. Despite the proliferation of digital learning tools, many current platforms lack the capacity to adapt dynamically to individual learner’s needs, thereby limiting their effectiveness. This study aims to develop and evaluate an intelligent adaptive learning platform leveraging artificial intelligence (AI) techniques to enhance personalized education. The primary objectives are to design an AI-driven framework capable of real-time assessment and tailored content delivery, to implement the platform using machine learning algorithms, and to empirically assess its impact on learner engagement and academic performance. The research adopts a mixed-methods approach, integrating quantitative and qualitative methods to provide a comprehensive evaluation of the platform’s effectiveness. The quantitative component involves a quasi-experimental design with a sample of 200 undergraduate students enrolled in introductory mathematics courses at a public university, randomly assigned to control and experimental groups. Data collection instruments include standardized academic assessments, engagement surveys, and system usage logs, while qualitative data are garnered through focus group discussions with learners and interviews with instructors to explore user experience and perceived benefits. The development phase employs a design science research methodology, with iterative prototyping and testing of AI modules including natural language processing, reinforcement learning, and predictive analytics. The analysis of quantitative data will utilize statistical techniques such as multivariate analysis of variance (MANOVA), regression analysis, and t-tests to determine the platform’s effectiveness on learning outcomes and engagement metrics. Qualitative data will be examined using thematic analysis to identify patterns related to user satisfaction, perceived personalization, and usability challenges. The development of the platform is grounded in established educational theories, including Vygotsky’s social constructivism and the Cognitive Load Theory, which inform the system’s adaptive mechanisms and content sequencing strategies. Expected findings include improved academic achievement among students using the adaptive platform, evidenced by higher test scores in comparison to traditional instruction, as well as increased learner engagement and motivation. The system’s ability to personalize content according to individual performance and learning preferences is anticipated to result in higher satisfaction levels among users. Furthermore, the empirical evaluation aims to demonstrate the feasibility of integrating various AI techniques—such as machine learning classifiers and predictive models—into a cohesive platform that responds to diverse learner profiles in real time. This research contributes novel insights into the design and implementation of AI-driven adaptive learning environments, filling existing gaps in the literature regarding practical integration of AI with pedagogical principles for scalable digital education solutions. It advances understanding of how intelligent algorithms can facilitate personalized formative assessments, dynamically adjust difficulty levels, and optimize learning pathways. The study’s theoretical contributions include validating the applicability of Vygotsky’s and Cognitive Load theories within AI-enabled adaptive systems. In conclusion, the study recommends the adoption of AI-driven adaptive platforms in higher education for fostering personalized learning experiences and improving academic outcomes. It suggests further research on scalability and long-term impacts across diverse educational contexts, as well as exploring ethical considerations related to data privacy and algorithmic bias. The findings aim to inform educational software developers, policymakers, and academic institutions seeking to leverage AI technologies to enhance pedagogical effectiveness by offering evidence-based, adaptable learning environments that cater to individual learner needs.