Since the Museum of Modern Art's opening in 1929, curators have used photography to document exhibitions before they close, but a new MoMA tool developed in partnership with Google Arts & Culture Lab is making those old exhibition photos a far more practical resource. Thanks to the advanced machine-learning algorithms developed for Google Arts & Culture, thousands of unlabeled photos have now been sorted and included in MoMA's online collection.
Though humans could identify the works in the photographs themselves, the issue is related to time. Identifying one image could take anywhere from a few seconds to hours of research, and multiplying that by 30,000 made the prospect of a human-run project impractical. With Google's algorithm, MoMA was able to nearly immediately identify over 20,000 of the artworks in the exhibition photos, creating an even more robust network for the museum's website.