Sbagliando s'impera

Un articolo di Ahmed Elgammal e Babak Saleh descrive un algoritmo per confrontare opere d’arte visiva e misurarne la “creatività” o quanto meno l’originalità e la capacità di influenzare altri artisti.

Qui trovate l’articolo originario in .pdf: Quantifying Creativity in Art Networks.

Qui, invece, un estratto dell’articolo comparso su Quartz l’11 giugno 2015: Picasso = Genius: This algorithm can judge “creativity” in art as well as the experts.

Art is seen as unquantifiable. Great paintings are creative forces that transcend their brush strokes, colors, and compositions. They can’t be reduced to mere data, analyzed, and ranked by their creativity. Two computer scientists at Rutgers University respectfully disagree.

Ahmed Elgammal and Babak Saleh created an algorithm that they say measures the originality and influence of artworks by using sophisticated visual analysis to compare each piece to older and newer artwork. They worked from the premise that the most creative art was that which broke most from the past, and then inspired the greatest visual shifts in the works that followed.

They did it by looking specifically at qualities such as texture, color, lines, movement, harmony, and balance. “These artistic concepts can, more or less, be quantified by today’s computer vision technology,” they write in their paper “Quantifying Creativity in Art Networks” (pdf).

Their experiment—which involved two datasets totalling more than 62,000 paintings—was entirely automated. They gave the computer no information about art history. Yet what they found was that their algorithm often came to same conclusions as art experts. “In most cases the results of the algorithm are pieces of art that art historians indeed highlight as innovative and influential,” the authors wrote.