This manuscript presents a collection of research conducted over the past twelve years during my tenure as an assistant Professor within the MOCAH team (Models and Computer Tools for Human-Centered Learning) at LIP6 (Sorbonne Université). These studies focus on modeling diagnostic and decision-making problems in the field of Technology-Enhanced Learning (TEL), with particular attention to adaptive feedback, adaptive scheduling, and student modeling. The models proposed have been tested in various contexts, such as serious games, intelligent tutoring systems, and programming learning platforms. My contributions more broadly fall within the field of Human-Centered Computing, which aims to design models and systems capable of making decisions or assisting humans in their decision-making processes.
The first chapter of this manuscript presents the computational models I have employed and adapted throughout my research. It highlights how these models and algorithms, drawn from computer science, can be applied — and systematically require adaptation —to address challenges specific to the field of human learning.
The following two chapters detail my main contributions, structured around two complementary axes: on the one hand, decision-making and decision support for selecting feedback or activities adapted to the learner; on the other hand, student modeling and assessment, in order to diagnose their knowledge level or predict their performance. This work takes place within complex learning environments, particularly those characterized by large state and action spaces, such as open-ended learning environments. In these open environments, my aim is to foster interaction and mutual enrichment between automated decision-making approaches and human expertise.
In conclusion, I summarize the research presented in the previous chapters to highlight the core of my scientific project. My research aims to deepen the understanding of human learning processes and adaptation mechanisms of TEL systems, by combining student modeling, human expertise, and artificial intelligence approaches. The main focus of my work is to place humans at the center of decision-making processes, by designing systems that are not only high-performing according to classical criteria (such as accuracy) but also acceptable, ethical, interpretable, and explainable.