An interesting summary in MIT Technology Review of some recent research done on creativity in historical art, creativity here being taken to mean novelty in imagery or content that had an influence on other– by definition less creative and more derivative– works by the same artist or by others. A machine vision algorithm analysed “classemes”: visual concepts which “can be low-level features such as color, texture, and so on, simple objects such as a house, a church or a haystack and much higher-level features such as walking, a dead body, and so on.”

Intriguingly, the algorithm is not restricted to figurative art and it can cope with abstraction and pop art, although at this stage they seem to be looking at painting. The software critic also tends to broadly agree with human assessments of the most influential works and artists even though it was not primed or biased in…