Workshop on Representation in/of AI with Fabian Offert

The Center for Culture and Technology, University of Southern Denmark and the research cluster “The Aesthetics of Biomachines and the Question of Life” (The Velux Foundations) invite you to join this exciting talk!

"This Is Your Brain on ImageNet": Embedding and Visual Epistemology – Fabian Offert (UCSB)

"Embedding" is one of the most important techniques in the machine learning toolbox. Polemically, in natural language processing and computer vision, any useful knowledge is embedded knowledge. While the technique itself is hardly more than an advanced form of compression, it is the universality of embeddings that renders them interesting from an epistemological perspective: universal faculties – such as "seeing" in the case of computer vision, which I will focus on in this talk – are extrapolated from particular datasets and represented in an exclusively relational manner. Exactly because of their universality, embeddings live on, sometimes way beyond the lifespan of the datasets that they represent. "Historical" deep convolutional neural network features, for instance, still inform the training of newer generative models by ways of perceptual distance metrics that determine the realism of generated images. By becoming just another part of the training pipeline, however, they cease to appear as distinctive epistemic structures. More importantly, this "historical opacity" of embeddings obfuscates what I propose to understand as a major epistemic shift: scientific knowledge, at least if it relates to the visual, is often generated with the help of cultural data. Embeddings, then, can be seen as a cultural technique, and as a trading zone that spans, surprisingly, not only branches of the natural sciences, but the sciences and the humanities.

Fabian Offert is Assistant Professor for the History and Theory of the Digital Humanities at the University of California, Santa Barbara, and principal investigator of the international research project "AI Forensics", funded by the Volkswagen Foundation. His research and teaching focuses on the question how machine learning models represent culture and what is at stake – aesthetically and politically – when they do. His current book project investigates "Machine Visual Culture" in the age of foundation models. https://zentralwerkstatt.org.

Zoom link.

Registration is not necessary.

Please contact: Naja Grundtmann for additional info.