Mr Giovanni Lion
Concept Formation in Computational Creativity: a Comparative Study of Algorithmic Approaches
Concept formation, Computational Creativity, Artificial Intelligence
The objective of the proposed research is to advance and extend the theory of concept formation in computational creativity. The primary focus is to understand how novel concepts come about in humans and whether it is possible for artificial agents to generate new concepts in a similar manner. I will adopt a transdisciplinary approach to this problem, investigating the interplay between existing theories developed in the fields of Philosophy, Cognitive Science, Design, Psychology and the state of the art in machine learning and artificial general intelligence. The aim is to formalize a theoretical framework and a methodology to evaluate the concept formation capabilities of multiple categories of artificial agents and then test the different algorithmic approaches against several problem spaces in multiple domains.
The primary research methodology adopted is a comparative analysis over the quantitative output of computer simulations. These computer simulations are formulated in such a way that they represent different theoretical frameworks in their corresponding algorithms. The simulation comprises of a diverse set of creative tasks that each artificial agent will try to accomplish. The output is compared across agents in order to collect insights on the validity of the agents’ underlying frameworks.
BSc – Bocconi University
MSc – Bocconi University
MSc – The Hong Kong Polytechnic University
Prof. dr. dr. Johan Hoorn (Main)
Prof. dr. Gino Yu (Co)
Specialization / Interests
Algorithmic Composition, Generative Art, Co-creative Systems, Generative Deep Learning
Date of Completion