Research Project: Training the Archive

Ludwig Forum Aachen, HMKV at the Dortmunder U and others

A research project on the development of algorithms for pattern recognition in big data research at the Ludwig Forum Aachen and the HMKV, Dortmund.

Visual Archives and Artificial Intelligence
Training the Archive explores the applicability of artificial intelligence to art and the curation of art. The 2020–2023 project is dedicated to visual archives and the question of how new contexts can be created in these collections using machine learning. The HMKV Hartware MedienKunstVerein, Dortmund, is cooperating with the Ludwig Forum für Internationale Kunst Aachen and the Visual Computing Institute at RWTH Aachen University.

The aim of Training the Archive is the research-based creation of software that enables curators and artists to gain new access to digitised image collections. The software is intended to automatically visualise similarities and differences between objects in digital collections and the connections between them.

Another main focus of the research project is the theoretical and discursive reflection of so-called 'artificial intelligence'. To this end, a number of specific formats are being developed and implemented at the HMKV. The publication format of the Working Papers reflects the current state of work and discussion within the research group. Ten working papers published in the course of the project make new knowledge available to the working group and also transport it to the outside world.

Mediation and Research
A series of video interviews with artists and curators forms another module. They differ in two different focuses. One part of the interviews collects empirical information, which is directly used in the development of the software prototype. Another part, which is mainly developed by HMKV, aims at the public perception of artistic and curatorial projects with 'artificial intelligence'.

Further Training
Another element of the research project consists of an internal further training, which is offered to the participating institutions. This is particularly concerned with conveying the complex knowledge about the technology, origins and social impact of 'artificial intelligence' to the institutions.

From Intelligence to Pattern Recognition
The term 'artificial intelligence' is addressed from a critical perspective in Training the Archive. We will ask whether the term 'intelligence' or 'cognition' is appropriate at all. Which terms could possibly be used in its place? For the intended project, for example, 'pattern recognition' is a more neutral term. This repositions the field of investigation of curating and artistic practice with 'artificial intelligence'. It no longer poses itself as a question of imagined intelligence, but as an examination of pattern recognition.

Post-AI Art
Post-Internet Art describes art that no longer understands the Internet as something new, but takes it as a given. In a similar sense, one could also speak of Post-AI Art. It is an art that does not understand 'artificial intelligence' as something new, but as commonplace. Artificial intelligence' applications involve pattern recognition in texts, speech and sound recordings and images, for example when digital cameras automatically recognise faces in order to bring them into focus, or in speech recognition and dictation functions.
The particular perspective that HMKV Hartware MedienKunstVerein can contribute stems from the understanding that pattern recognition is increasingly embedded in existing software applications. This is most obvious, for example, in new filters for the graphics editing software Photoshop that automate object recognition, or remove skin wrinkles. Less obvious, but nevertheless powerful, is this in business software, such as SAP or Salesforce, in logistics software, timetable control in public transport and the like. In this sense, post-AI art is not to be read solely as an aesthetic endeavour, but in an infrastructural, political and social sense.

Training the Archive
From this perspective, the envisaged software for curators and artists to access archives through pattern recognition is becoming increasingly complex. The project partners are therefore con-tributing various expertises and discussing them in project meetings and working papers: The Ludwig Forum für Internationale Kunst Aachen is designing a software application for collection access based on computer vision and pattern recognition. The technical partner Visual Computing Institute of RWTH Aachen University is involved in the implementation. The HMKV contributes discursive expertise to the project, which discusses the respective stages of work within the project and anchors the question of Training the Archive in the public sphere.
The results of the research project will be published continuously on the websites of the HMKV, the Ludwig Forum für Internationale Kunst Aachen and GitHub. For a professional exchange, a symposium with international experts is planned for 2022. The challenge of Training the Archive lies in this complexity and its communication.


Ludwig Forum Aachen and HMKV Hartware MedienKunstVerein


Funded by:

Digital Culture Programme of the Kulturstiftung des Bundes (German Federal Cultural Foundation)


Funded by:

the Beauftragte der Bundesregierung für Kultur und Medien (Federal Government Commissioner for Culture and the Media).

Video documentation on the “Training the Archive” conference

Inke Arns, Sybille Krämer & Yvonne Zindel (Panel discussion, German language only) – Culture between Humans and Machines: Who Curates Whom?

Sybille Krämer (Keynote, German Language only) – Digital Culture. Fundamentals and Ambivalences of Algorithmisation

Dominik Bönisch (Lecture, German language only) – How to Train the Curator’s Machine

More videos from the conference can be found on the Training The Archive blog: here.


Video Interviews

Interview #5 with Elisa Giardina Papa on »Emotion and AI«

You can find more information about the interviews on the Ludwigforum's Training the Archive blog.

Interview #4 with Matteo Pasquinelli on »Exploring the economic and social roots of AI«

You can find more information about the interviews on the Ludwigforum's Training the Archive blog.

Interview #3 with Nick Couldry & Ulises Mejias on »Data Colonialism«

You can find more information about the interviews on the Ludwigforum's Training the Archive blog.

Interview #2 with Adam Harvey on »Face Recognition Datasets«

You can find more information about the interviews on the Ludwigforum's Training the Archive blog.

Interview #1 with Magda Tyzlik-Carver on »Curating Data«

You can find more information about the interviews on the Ludwigforum's Training the Archive blog.



Francis Hunger
Artistic-scientific Assistant "Training the Archive"



The latest information and reports about the project can be found on the official blog:

Go back