Emotion-Aware Course Personalization in Virtual Classrooms via Multi-Modal Fusion Learning Systems

Authors:
Publisher website
Google Scholar
Abstract
Emotion-aware course personalization in virtual classrooms aims to enhance the learning experience by dynamically adapting content based on students’ emotional and cognitive states. This approach leverages AI to create responsive and human-centric e-learning environments. Existing virtual learning systems often rely solely on cognitive metrics, such as quiz scores or activity logs, failing to consider emotional cues like confusion, frustration, or disengagement. As a result, learners experience reduced motivation, lower retention, and poor personalization. To address these challenges, we propose a novel Multi-Modal Deep Attention Fusion (MMDAF) framework that integrates data from facial expressions, voice tone, and textual interactions. By employing deep learning and attention mechanisms, MMDAF selectively weights and fuses inputs to more accurately detect learners’ emotions in real-time. The detected emotional states are used to personalize the learning path by modifying content delivery, pacing, or interactive support. For instance, confusion can trigger the use of simplified explanations or visual aids, while boredom may lead to the implementation of gamified modules. Experimental evaluations in simulated virtual classrooms of learner engagement, emotional satisfaction, and academic performance compared to traditional models.

This publication has 39 references indexed in Scilit: