Enhancing personalized learning with deep learning in Saudi Arabian universities

Abstract

This study explores the use of deep learning methods in personalized learning environments to improve educational outcomes. We collaborated with four major universities in Saudi Arabia and used data from the Blackboard Learning Management System to gather insights on various personalized learning approaches. This helped us develop a flexible model that is suitable for different learning environments, guided by the VARK model. We used a hybrid deep learning model combining Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and Recurrent Neural Networks (RNNs) to classify students based on their learning preferences and engagement patterns. Our analysis showed significant improvements in student motivation and engagement with personalized learning materials. The results indicated high satisfaction levels among students and faculty, with the model achieving 85% accuracy in predicting student engagement and recommending personalized learning paths. Training the model on a dataset of 10,000 student records took about 12 hours, with 80% GPU utilization during training and 30% during inference. Precision and recall rates were 82% and 88%, respectively, with an F1-score of 0.85. Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) were low at 0.15 and 0.20, respectively. Integrating deep learning methods into personalized learning environments represents a significant shift in education, enabling educators to enhance student engagement and performance effectively. Collaboration with faculty members highlights the importance of interdisciplinary approaches in advancing educational technology and pedagogy, ensuring stakeholder satisfaction and success.