Multimodal AI Models for Personalized Learning in Children with Special Needs: Integrating Computer Vision, NLP, and Human-AI Interaction

This paper presents EduSense-AI, a novel multimodal artificial intelligence framework designed to deliver personalized learning experiences for children with special needs. By integrating computer vision, natural language processing, and speech emotion recognition, the system dynamically adapts educational content based on learners’ cognitive and emotional states. A pilot study involving ۱۲ children over four weeks demonstrated significant improvements in engagement, retention, and emotional stability. The architecture leverages transformer-based models and reinforcement learning to optimize content delivery in real time. Beyond technical innovation, the paper critically examines ethical considerations such as data privacy, algorithmic bias, and educational equity. The proposed framework aligns with the vision of inclusive AI and offers actionable recommendations for integrating such systems into national educational infrastructures. Drawing on recent research, the article synthesizes insights into how AI-driven multimodal systems enhance inclusivity, adaptive learning, and emotional engagement, while addressing ethical and technological challenges in special education.

کلیدواژه ها:

Multimodal Artificial Intelligence ، Personalized Learning ، Special Education ، Computer Vision ، Natural Language Processing ، Emotion Recognition ، Educational Equity ، Human-AI Interaction ، Transformer Models ، Ethical AI