Training the Archive

Facebook Instagram Github

Prototype for clustering museum collection data published

Scatterplot of the walk through the high-dimensional space from one artwork to another, the direction of the path is illustrated by arrows. All images are open-source data from the SMK - Statens Museum for Kunst, Copenhagen.

The first Prototype in the research project »Training the Archive« has been published. It makes it possible to use Clustering of image data from a museum collection to identify interesting visual connections between artworks. Subsequently, humans are to be integrated again as experts by training the knowledge of curators to the neural network used for image recognition. This is intended to make the clusters formed more dynamic and personalised.

The Repository with all the notebooks and the instructions can be found here:

»Training the Archive« (2020–2023) is a research project that explores the possibilities and risks of AI in relation to the automated structuring of museum collection data to support curatorial practice and artistic production.

Collaborative Partner:
Digital Partner:
Funded by the Programme:
Funded by:
Github Facebook Instagram