Artificial Convergent Evolution
This artwork is a visualization of a novel AI-based convergent evolution process that is inspired by convergent evolution in nature. It is a result of an experiment that investigated the aesthetics that emerge when generative models are isolated and then pushed towards each other to converge.
Convergent evolution is a phenomenon in which organisms that are not closely related independently evolve similar traits in response to similar environmental pressures. For example, human eyes and octopus eyes evolved to be similar in structure and function but they were not inherited from a common ancestor. Just like different environmental conditions may cause species to branch off, similar environmental conditions sometimes lead to convergence into a similar appearance despite taking very different paths.
This artwork was created using a custom-designed AI art model that emulated this process: isolation, branching off, and convergence. The model was first trained on a large, generic image dataset. Then, copies of the model were further trained on separate datasets in isolation. Since the copies only saw disjoint sets of data, they evolved to specialize on the limited set of images they observed. After they branched off, the copies were pushed towards each other to converge into similar evolutionary paths.
This novel artificial convergent evolution process helped generative models diverge from their original training sets while converging towards similar aesthetics, resulting in visually distinct and aesthetically pleasing images.
Artsy Editorial: Why These 8 Artists Are Making NFTs