Title: Design of automatic generation model for teaching music based on user differences and transfer learning

Authors: Xingyu Huo

Addresses: School of Theatre, Film and Television, Communication University of China, Beijing, China

Abstract: The music industry's development and innovation are vital for expanding music education channels and enhancing cultural and artistic innovation. To achieve digital and intelligent music teaching tailored to students' needs, this study developed a music automatic generation model using pre-training and transfer learning. By integrating semantic analysis and user differences, the model offers enhanced personalised features. Performance analysis indicates high precision and recall rates at 91.49% and 88.42% respectively, with average errors below 0.550 and 0.450. The model's quality has significantly improved, providing personalised music that aligns with user needs. Through subjective and objective evaluations, the model's generated music demonstrated good quality. Overall, this model boosts music creation efficiency, fosters innovative music art development, expands music education outlets, enhances beginners' understanding of music creation processes, and popularises music education.

Keywords: semantic analysis; attention mechanism; intelligent teaching; automatic music generation model; transfer learning; user differences.

DOI: 10.1504/IJES.2024.143770

International Journal of Embedded Systems, 2024 Vol.17 No.1/2, pp.62 - 72

Received: 12 Jan 2024
Accepted: 02 Jun 2024
Published online: 06 Jan 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article