Deep Learning and Transformer Models for Educational Chatbots
Siva Subramanian R, G. Anurekha, C. Gomathi, Sucharita Saha, A. Rajalakshmi, J. Elavarasi, M. Ezhilvendan · Advances in computational intelligence and robotics book series · 2025
The integration of deep learning and transformer models has revolutionized educational chatbots, enhancing the capabilities of intelligent tutoring systems (ITS). This chapter explores the evolution of AI-driven chatbots, focusing on Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNNs), and reinforcement learning for adaptive learning experiences. Additionally, transformer-based models enable superior natural language understanding (NLU), improving chatbot interactions in education. Key applications include personalized tutoring, automated grading, multilingual support, and student progress tracking. However, challenges such as data bias, explainability, privacy concerns, and scalability persist. Future advancements in few-shot learning, knowledge graphs, sentiment-aware tutoring, and AI-human collaboration will further enhance AI-powered education. This chapter highlights emerging trends, challenges, and research directions in AI-driven intelligent tutoring systems.