Application of Artificial Neural Network in Chinese Folk Music

摘要

In recent years, artificial neural network have been widely used in music application. Algorithm can compose music that is comparable to human performances. However, few people use algorithm to compose Chinese folk music. Applying algorithm to ethnic music composition is beneficial to its development. In this paper, we present a Chinese folk music sequence learner based on the superimposed LSTM neural network model. By considering the correlation between pentatonic scales, we create ethnic music that has more sound structures. We also conduct experiments to evaluate the quality of our music generation.