Artificial intelligence-based emotion recognition application of English teaching in smart learning

Abstract

With the rise of the intelligent era, innovative learning has gained increasing attention, particularly regarding students’ needs. While English teaching has moved away from the ‘dumb English’ approach, focusing more on integrating listening, speaking, reading, and writing, many still view English as a subject rather than a language, affecting teaching effectiveness. Due to spatial and temporal limitations, emotional interaction between teachers and students is lacking. This study explores an AI-supported emotion recognition teaching model, integrating relevance, originality, and impact (ROI) theory with innovative English education. An echo state network was constructed, and the algorithm was optimised. Emotion classification and speech signal preprocessing were implemented. Experimental results show improvements in students’ performance in vocabulary (3.8%), listening (4.5%), reading (5.9%), and speaking (7.1%) compared to traditional methods, enhancing smart learning quality and classroom interaction.