Computational Creativity and Aesthetic Measure

As a step towards formalizing the cognitive process and interpretation of creativity, we are developing a theory and computational model of creative systems that aims to better define the evaluation of creative systems and processes. We extend previous research, from Boden() and Wiggins() of defining a formal framework for describing and comparing creative systems by taking into account the judgment of artefacts that they generate, with respect to an observer. Prior work on formalizing creativity has mainly focused on the generating system with respect to the evaluation of created artefacts. This project models the judgment of creativity through explicit representation  of observer systems.

We have shown how this extended theory can support the interpretation of MacGyver problems—problems defined in the cognitive systems research community as classical planning problems designed to elucidate the cognitive process of creative problem solving.

This theory also allows for the development of an evaluation framework of creative systems by putting forth general criteria of subjective evaluation based on initial empirical work regarding aesthetic measure across artistic domains. This is done by using statistical methods to model individual and group preferences over creative artefacts with respect to their judgement by different observers. These pilot experimental results model abstract metrics of balance, symmetry, and readability/coherence that are used in informal human subjective evaluation of judgement across artistic and creative domains.

More details in this working paper — [ObserverTheoryOfCreativity]