Representação multiparadigma de conhecimento musical utilizando programação lógica indutiva

Abstract

Knowledge representation process is an essential matter regarding Computer Music systems. Methods have been applied in order to provide computers with the capability to generate conclusions based on experience in specialized domains. Inductive Logic Programming is a research field which combines concepts of Logic Programming and Machine Learning. Due to its declarative feature, both acquired and produced knowledge can be presented to not-expert users in a naturally understandable way. This work deals with Music Knowledge Representation from the perspective of multi- paradigm programming, using Inductive Logic Programming technique and including the development of the knowledge-based music system Fraseado. Finally, a method for the evaluation of algorithmic composition systems - the Expanded Turing Test - is presented.