Abstract
In recent times, characterized by the rapid advancement of science and technology, the educational system has continuously evolved. Within this modern educational landscape, Science, Technology, Engineering, Arts, and Mathematics (STEAM) education has emerged as a prominent pedagogical paradigm, gaining substantial popularity in college-level instruction and capturing widespread societal attention. Notably, the cultivation of audio-visual aesthetic proficiency occupies a central role within this educational approach, prioritizing the enhancement of aesthetic sensibilities. By ingeniously amalgamating scientific knowledge with emotional expression, this research assumes a crucial facet in the holistic development of individuals. The research aims to explore the cultivation of students' audio-visual aesthetic abilities in university-level vocal music education by integrating deep learning and STEAM education principles. Drawing upon an extensive review of relevant literature, this research synthesizes the principles of STEAM education with those of deep learning, while considering the current cultural and societal context and the distinct realities faced by contemporary college students. Consequently, this research posits a novel conceptual framework for curriculum design and proposes a three-stage teaching process model. To substantiate the efficacy of this innovative educational model, an empirical investigation employing a questionnaire survey is conducted to assess its teaching effectiveness, confirming the marked superiority of this pioneering pedagogical approach. The results demonstrate that the new teaching model has led to notable enhancements in students' audio-visual aesthetic abilities, self-confidence in learning, and learning efficiency. Additionally, compared to traditional educational methods, the curriculum primarily, which focused on STEAM education with the project as its core, emphasizes the logic of the learning process and its connection with other disciplines. In conclusion, the three-stage educational model combining STEAM education and deep learning fully considers students' learning situations and utilizes the analytical capabilities of computers for educational purposes. This learner-centric approach significantly augments teaching efficiency and flexibility. Finally, the research concludes by summarizing its contributions and limitations, offering practical recommendations for the field. This research provides new insights and references for the practice and improvement of audio-visual aesthetic education in higher education institutions.