Digital Signal Processing and Artificial Intelligence for Automatic Learning TaoZhang 3 3 2024 https://doi.org/10.6025/dspaial/2024/3/3/83-89 https://www.dline.info/dspai/fulltext/v3n3/dspaiv3n3_1.pdf In recent years, the advancement of artificial intelligence technology has become a significant driving force for digital music creation, attracting widespread attention. Compared to the high cost and uncertain timing of manual creation, the extreme cost-effectiveness of AI music can effectively help developers reduce costs and increase efficiency. Therefore, the automation of digital music creation has broad applications and significance. The recursive neural network is an artificial intelligence algorithm based on neural networks, mainly with a loop structure, that can adaptively process sequence data. In a digital music creation system, recursive neural networks can predict and generate note sequences, thereby achieving automated music creation. Various factors, such as the quality of training data, the settings of model parameters, and the stylistic characteristics of the music influence the predictive results of recursive neural networks. Therefore, meticulous parameter tuning and optimization need to be conducted for different application scenarios.