Abstract:
Machine Learning has become increasingly important for the processing and visualisation of digitised collections in art museums. With the use of artificial intelligence, various models can be applied in order to take a new look at digital objects and gain fresh insights from the available information. The research project Training the Archive seeks to use computational methods to reveal connections and links between artworks, thus supporting curators in their practice. Since curatorial research is closely entangled with technologies such as search engines, collaborative approaches based on a human-machine interaction are becoming increasingly common. In this paper, different possibilities are discussed for establishing a similar form of collaboration between curators and Machine Learning models. Three suggestions are made, ranging from current prototypes to future possibilities. In each case, the aim is to translate expert knowledge into a meaningful tool – introduced as the Curator’s Machine.