Chord Generation Based on Transformer and Color Harmony Theory

Artificial intelligence has been widely used in the field of music and achieved good results. As an important concept in music, chord generation is one of the basic tasks in composition. The application of machine learning in music is of great significance to human’s understanding of artificial intelligence and music, and artificial intelligence technology can also enrich people’s research in music. In this paper, a chord generation model is constructed based on Transformer to further dig into the internal structure of chords and consider chords from the aspect of chord notes, hoping to generate more delicate and flexible chords. At the same time, combined with the color harmony theory, the evaluation method and mapping method meeting the rules of music theory are constructed, hoping to get better output effect through a simple model, and the template mapping method is used to improve the actual effect and extensibility of the model.