Teaching activity design and psychological practice of music majors under a convolutional neural network and transformer module

The purpose of this study is to explore a new mode of teaching activity design and psychological practice for music majors supported by the concept of STEAM education. In this study, a personalized recommendation system based on a convolutional neural network (CNN) and a Transformer module is designed and implemented. The STEAM education concept is further introduced, which is combined with a personalized recommendation system to design music teaching activities to improve students' learning effect and comprehensive quality. Finally, the performance of the model is evaluated experimentally. The results show that the accuracy, recall, and F1 values of the personalized recommendation system designed here reach 95.76%, 90.28%, and 92.41%, respectively, which is more than 4% higher than other reference models. And in the user behavior indicators, the number of clicks, browsing time, and comments has increased by more than 10%, which reveals the remarkable advantages of the proposed algorithm in improving teaching effect and learning experience. Therefore, the proposed CNN-Transformer algorithm shows obvious advantages in prediction accuracy and user participation, which provides new empirical support for the application of educational technology in music education.