Concept Formation in Computational Creativity: A Comparative Study of Algorithmic Approaches
Abstract
The doctoral thesis presented in this paper addresses the latest advancements in machine learning architectures for creative artifact generation through the lenses of Design, Philosophy and Cognitive Science. The research adopts a trans-disciplinary approach, looking for opportunities to decompartmentalize knowledge and formalize efficient guidelines that facilitate adoption and operation of such technologies. Its main objective is to uncover the hidden assumptions embedded in these tools and formalize a theoretical framework able to describe the different concept ontologies employed during human-machine collaboration and their inter-operability. Three studies will be conducted to validate the framework, each addressing a specific domain (music, language, visual).
