Student name: Gadea Lucas Pérez

RUN-EU institution: University of Burgos, Spain

Abstract

Learning is not a one-size-fits-all process. Everyone absorbs, processes, and retains information in a different way based on prior knowledge, cognitive abilities, and even emotional states. Traditional learning and training methods often fail to accommodate these differences, leading to frustration for some and overconfidence for others. This is where intelligent self-adaptive systems come into play, providing personalised training that can adjust to the learner’s needs in real time.

In this presentation, we discuss how adaptive learning, combined with immersive Virtual Reality (iVR) and physiological sensors, can greatly improve skill development. Hence, by continuously assessing user performance and physiological signals —such as eyetracking, heart rate, electrodermal activity—these systems can dynamically modify the training scenarios to maximise users’ engagement and effectiveness.

The goal is to combine this VR environments with Machine Learning (ML) techniques. Thus, we can simulate stressful situations in a realistic and controlled way, while offering tailored feedback that supports all learners: both novices and experts. For instance, the experienced learners may operate on «autopilot» and often underestimate challenges. In this situation, adaptive systems can introduce variability to refine their skills. At the same time, new users benefit from guided support, reducing frustration and enhancing their learning curve.

Moreover, Transfer Learning (TL) principles enable the application of acquired knowledge across different contexts, improving scalability and reducing resource demands. Still, a lot of challenges remain, such as ensuring real-time adaptation, minimising reliance on hardware, or evaluating the system’s effectiveness in diverse learning environments.

By harnessing adaptive systems in immersive VR settings, we move toward a future where learning is no longer rigid but dynamic, responsive, and truly personalised.