“Towards a Theory of Machine Learning and the Cinematic Image” presented by Lyons

Symposium:

• ISEA2022: 27th International Symposium on Electronic Art
• More presentations from ISEA2022:  

• Algorithms, Machine Learning and Audiovisual Heritage

• Towards a Theory of Machine Learning and the Cinematic Image

• Owen Lyons

Venue(s):

• CCCB: Centre de Cultura Contemporània de Barcelona, Barcelona, Spain

• This paper addresses state-of-the art and prehistory of machine learning technologies as they pertain to the moving image. It discusses the challenges and potential hazards of these technologies, before showing how artists and image-makers are using these techniques to create a new relationship to the space of the cinematic in the age of big data. This paper strives to interrogate and overcome technologically deterministic accounts of this new technology. Accordingly, it situates examples within the discursive fields from which they emerge and favors diachrony over synchrony in outlining four clusters of media-archaeological inquiry that intersect in the application of machine learning techniques to the moving image. I believe that this approach is a necessary intervention into the hegemonic discourse on machine learning that has emerged from the technology sector which obscured deep ideological problems within the algorithms themselves, and also overlooks the potential of these technologies.

Keywords:

• Machine Learning
• artificial intelligence
• Cinema
• cultural analytics
• surrealism

Category:

• Long Paper
• Paper