What are you looking for?


The increase in computing capacity and the use of convolutional neural networks (CNN – A kind of deep neural network commonly applied in image analysis) have favoured and sped up the development of new scenarios that allow access to large amounts of images.

The exhibitions “Revolutionary Arkive: The eye-machine is watching” (2018) and “Mnemosyne 2.0: Computational cartographies of visual memory” (2019) by Pilar Rosado point to the intimate interrelation between the production of images, iconicity and culture. The genealogy of these interests dates back to the historian of art and culture Aby Warburg (1866-1929) who, with his Atlas Mnemosyne, made up of a collection of images and little text, sought to narrate the history of the memory of European civilization, considering that the images by themselves and in their mutual relationship generate a space for thought. In this sense, the installations can be seen as an exercise in repositioning Warburg’s ideas using 21st-century tools (CNN, ResNet-50 and algorithms such as t-SNE) to broaden the debate on the notions of collective visual heritage and its role in the culture.

In the exhibitions presented, the possibility of covering the total visual content of a collection allows us to consider information that we would not have access to by analysing only the parts, in addition, a co-creation dialogue is established in which the visual stereotypes of the author with the visual patterns detected by the algorithm in the large groups of photographs analysed. These methodologies, which anticipate the photographic archives of the future, open up new avenues for dialogue with the past and promise to illuminate many aspects of the history and evolution of images.

Pilar Rosado

Artist, teacher and researcher. With a degree in Biology and a doctorate in Fine Arts, she has published several essays on the application of artificial vision models for the analysis of large collections of abstract art images, providing alternative points of view for reflection on the conventions of our gaze. In her artistic practice, she explores political issues that can be approached from the image and that involve machine learning technologies, such as the management of information in the visual archives of the future, the revision of collective memory or artificial creativity.

Web: www.pilarrosado.eu

Referral links:

Computer Vision Models to Categorize Art Collections According to the Visual Content: A New Approach to the Abstract Art of Antoni Tàpies.


Panoramic views on the collective visual heritage through convolutional neural networks. The exhibitions Revolutionary Arkive and Mnemosyne 2.0 by Pilar Rosado.


Mnemosyne by Pilar Rosado from Pilar Rosado on Vimeo.