Implementing Adaptive Learning Systems
Tong San-hong, Wei YaoYang, Wei Li, Pei Jun, Uzma Sarwar · Advances in computational intelligence and robotics book series · 2025
The artificial intelligence (AI) revolution has dramatically transformed education, with the possibility of personalizing learning through adaptive systems within reach. Despite these promises, classroom adoption of AI-based systems is still difficult and faces various barriers, including resistance from educators, issues of privacy, and low scalability. This study addresses said challenges by developing a new hybrid adaptive learning framework combining Collaborative Filtering (CF) and transformer-based BERT models. While CF makes recommendations based on the behavior of the students, BERT improves personalization with context-aware and semantic insights, capturing the emotional and cognitive nuances. The framework was validated with 300 students and 50 instructors, showing a 40.32% improvement in quiz scores, a 30.91% increase in resource engagement, and a 32.35% rise in recommendation accuracy. This proves hybrid system overcomes many of the obstacles to adoption, presenting a scalable, ethical solution in personalized education within and beyond traditional teaching practice