Symbolic Music

Symbolic music research focuses on representing and processing music using discrete symbols, enabling computational analysis and generation. Current research emphasizes developing and applying deep learning models, particularly transformers and graph neural networks, to tasks such as music transcription, generation, and style transfer, often employing novel tokenization strategies to improve model performance. This field is significant for advancing music information retrieval, facilitating music education through automated difficulty assessment, and enabling new forms of human-computer musical interaction. The availability of large, open-source datasets is also a key area of ongoing development.