Expansion on description-based design of melodies

Abstract

This work-in-progress paper describes attempted improvements on Pachet's Description-Based Design (DBD), a system that uses machine learning to generate melodies. We discuss in depth both Description-Based Design and our extensions to Pachet's original approach. We also present a user study in which users had some success in transforming melodies and describe the implications of these results for future work. Copyright © 2013, Association for the Advancement of Artificial Intelligence. All rights reserved.