Keynote David Beer - Algorithmic thinking, neural networks and the presences of nonknowledge
Sprache des Titels:
Englisch
Original Kurzfassung:
Keynote by David Beer (University of York).
Algorithmic thinking is full of tensions. Framed within this wider observation, this talk will focus upon one instance of these tensions. Through their ?regimes of recognition?, as Louise Amoore has described them, neural networks are active in producing tensions between the knowable and the unknowable. This talk explores the role of unexplainability and unknowability in neural networks and, therefore, in emergent forms of artificial intelligence. In particular it will looks at how unexplainability is something that is actively pursued in the advancement of neural networks. The authenticity and validity of neural networks is based upon the presence of unknowability. Taking Georges Bataille?s notion of ?nonknowledge? as a conceptual touchstone, the talk will explore how nonknowledge is essential to the ideals, development and mutation of neural networks. Examining neural networks and their integrated development of depths and layers, the talk will look at how the presences of nonknowledge are to be found in forms neural networks take and in the discourse used to expound them.
David Beer is Professor of Sociology at the University of York. His previous books include The Data Gaze, Metric Power and Social Media and the Automatic Production of Memory (written with Ben Jacobsen). His new book, The Tensions of Algorithmic Thinking, will be published by Bristol University Press later in 2022. More information about his work can be found on his website.
Forschungs-