Machine Learning Methods for Artificial Musicality
Music representation learning is a type of machine learning methods that automatically extract music features, which can help downstream tasks such as music generation to achieve better performance. This PhD project will investigate music structure representation learning. The project will conduct research in the field of music structure analysis, design a reasonable music structure generation algorithm and use it to generate structured music with various changes in music. The project will include three applications based on music structure representation learning, including automatic video editing, music variation generation, and music summarization.
C4DM theme affiliation: