The objective of my research is to develop theoretical advances in artificial intelligence, signal processing and interface design that enable the key technologies required to automatically find, label and manipulate important structures in audio, with a focus on music as a problem domain