Evidence of the Transmutation of Creative Elements Using a Computational Creativity Approach

Abstract

Computational creativity is a relatively new subfield of artificial intelligence (AI). In short, it focuses on the ability of computers to generate objects that most humans with sufficient domain knowledge would consider having creative value, e.g., music, art, puzzles. In this article, the author presents rare evidence of the transmutation of creative elements or properties from one domain to another using a computational creativity approach, i.e., the Digital Synaptic Neural Substrate (DSNS). Specifically, in the domain of chess problem composition. This evidence suggests that creative elements, typically in the form of fragmented information or data, can indeed be transferred between domains. For instance, by drawing upon the characteristics of seemingly unrelated images (e.g., paintings, photographs) and sequences taken from chess games, valid chess problems or puzzles can be automatically generated. It is generally difficult to fully trace this kind of transmutation and simultaneously rule out algorithmic repetition or chance. However, there was at least one distinct example that left little room for contention. The author presents arguments for this "creative transmutation," additional supportive examples that were discovered, and likely directions for future work.