Computer Science
Probing for consciousness in machines
M. Immertreu, A. Schilling, et al.
This study investigates whether artificial agents can develop Damasio-style core consciousness by forming self and world models as byproducts of reinforcement learning in a video-game environment. Using feedforward probe classifiers on network activations, the authors find rudimentary self/world models, suggesting a pathway toward machine consciousness. This research was conducted by Mathis Immertreu, Achim Schilling, Andreas Maier, and Patrick Krauss.
Related Publications
Explore these studies to deepen your understanding of the subject.
