Music Understanding

Music understanding research aims to computationally replicate human comprehension of music, enabling applications like music generation, analysis, and information retrieval. Current efforts heavily utilize foundation models, including large language models and diffusion models, often incorporating multimodal approaches that integrate audio and textual data, and leveraging graph-based representations to capture musical structure. This research is significant for advancing AI capabilities in music and fostering human-computer collaboration in creative processes, as well as providing new tools for musicologists and music theorists.