A music knowledge representation system combining symbolic and analogic approaches

This paper describes the artificial intelligence (AI) methodologies and the system's architecture of a prototypical intelligent composer's assistant, called HARP (Hybrid Action Representation and Planning). The model is hybrid, since it stands on both symbolic (structured multiple inheritance semantic nets, SI-Nets) and analogic knowledge representation paradigms. Composition processes are modeled in this framework as plans to be accomplished to reach given goals. The paper focusses on the symbolic part of the system: how to store the analogic descriptions of processes in a knowledge base, and how to use them in a music composition environment.