System for Representing Musical Knowledge

Title: System for Representing Musical Knowledge

Research Question: Can a hybrid system that integrates sub-symbolic and symbolic processes effectively represent and model the knowledge involved in a complex musical activity?

Methodology: The researchers developed a hybrid system composed of a connectionist module and an agent-based module to combine sub-symbolic and symbolic levels. They chose first species two-part counterpoint melodies as a case study. The system captured different aspects of the musical activity by computing and applying three distinct functions: rules, fuzzy concepts, and learning.

Results: The researchers demonstrated the feasibility of modeling the knowledge involved in a complex musical activity by integrating sub-symbolic and symbolic processes. Their hybrid system successfully represented musical knowledge and enabled the musician to encode his knowledge, intuitions, and aesthetic taste into different modules.

Implications: This research has significant implications for the field of computer music and artificial intelligence. It presents a new approach for composing polyphonic music and offers insights into knowledge representation in complex musical tasks. The hybrid system can serve as an interactive tool for composers or performers, demonstrating its potential practical applications.

Link to Article: https://arxiv.org/abs/0203021v1 Authors: arXiv ID: 0203021v1