Research on the Application System of Deep Learning in English Remote Network Classroom

This study proposes an English remote classroom application system based on deep learning, which combines students' learning behavior data, video data and voice data for multimodal analysis. Specifically, the system uses the GNN model to model students' learning behavior and interaction relationships to reveal the interaction patterns between students; the self-attention mechanism is used to analyze students' classroom performance, dynamically focus on important features in the learning process, and provide personalized learning resource recommendations. Through experiments on the learning data of 1,000 students, the system can predict students' learning progress in real time, evaluate their emotional state, and generate personalized feedback and teaching adjustments. The experimental results show that the learning progress prediction accuracy based on the system reaches <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 3. 4} \%$</tex>, and the sentiment analysis accuracy reaches 91.5 %, which is significantly improved compared with the traditional evaluation method (accuracy is about <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 2 \%} \%$</tex>). In addition, the system can also dynamically adjust the teaching content according to the students' emotional state and learning progress, improve classroom interactivity, and further optimize the learning effect of distance education.