Are there such things as a truly unique piece of artwork? After all, artists derive their inspiration from all kinds of things, so sometimes subconsciously, they might have created a piece of artwork that could resemble something else without them knowing, and it is this connection that MIT’s AI is able to sniff out.

Developed by researchers from MIT’s Computer Science and Artificial Intelligence Laboratory and Microsoft, they have created an algorithm that is capable of discovering potential hidden connections between art pieces that might not have otherwise been discovered by humans just by looking at it.

The AI dubbed MosAIc was inspired by a special exhibit called “Rembrandt and Velazquez” at the Rijksmuseum. By using deep networks, it attempts to find analogous works from different cultures and artists. One of the examples was Francisco de Zurbarán’s “The Martyrdom of Saint Serapion” and Jan Asselijn’s “The Threatened Swan” which despite featuring very different subjects, were found to portray “profound altruism with an eerie visual resemblance”.

According to lead author Mark Hamilton, “These two artists did not have a correspondence or meet each other during their lives, yet their paintings hinted at a rich, latent structure that underlies both of their works.” The AI also goes beyond just recognizing colors, but also the theme and the meaning behind the artworks.

Hamilton adds, “Going forward, we hope this work inspires others to think about how tools from information retrieval can help other fields like the arts, humanities, social science, and medicine. These fields are rich with information that has never been processed with these techniques and can be a source for great inspiration for both computer scientists and domain experts. This work can be expanded in terms of new datasets, new types of queries, and new ways to understand the connections between works.”

Filed in General. Read more about , and . Source:

Discover more from Ubergizmo

Subscribe now to keep reading and get access to the full archive.

Continue reading