Art and science are often not connected, but researchers at MIT and Microsoft have wanted to counter this general belief by creating an algorithm that can analyze works in many ways and search for links.

You would have to look very carefully or have the knowledge of an art expert to see similarities in these two pictorial works. That is the achievement attributed to the new algorithm of the Laboratory of Computer Science and Artificial Intelligence (CSAIL) from MIT.

“The Martyrdom of Saint Serapion” by Francisco de Zurbarán and “The Swan Threatened” by Jan Asselijn, are “two works that portray scenes of deep altruism with a strange visual similarity” explain the researchers of this project. So his MosAIc program has linked both works.

Mark Hamilton, a CSAIL PhD student, explains in the MIT information that developing this algorithm was quite a challenge. Not only did they want recognize similar images for color or style but also for the meaning and theme of the work, something really complex, even for humans.

Kurt Luther, a professor at Virginia Tech University, has created a website that allows you to analyze old photographs to determine who the person who appeared in them was, and therefore be able to give those images the context they lack.

“It is impossible even for the most experienced art critics to take millions of paintings over thousands of years and to find unexpected parallels in themes, motifs and visual styles,” they say from MIT. This is one of the many objectives that this artificial intelligence has that they have trained with 5,000 images.

MosAIc seeks to help us explore existing art and facilitate access to millions of works of art, for experts or students, for example. User enters image and MosAIc algorithm finds similar artwork. Researchers may ask you for an instrument close to a blue and white dress, the algorithm responded with Japanese and Dutch works that had that detail in common.

Hamilton adds that he hopes the work started with MosAIc can expand to other fields such as humanities, social sciences and medicine. “These fields are rich in information that has never been processed with these techniques and can be a source of great inspiration for both computer scientists and experts in those fields,” he says.