Wednesday, 19 June 2013, 7:30pm - 8:00pm
Daniel Polani, Reader in Artificial Life, Adaptive Systems & Algorithms Research, University of Hertfordshire.
Recent studies indicate that Shannon information can act as a “currency” organizing cognitive load in agents via various optimization principles, such as a principle of “informational economy”.
This provides formulations for biologically plausible cognitive models of embodied agents, i.e. organisms or robots. Different from other bio-mimicking models (e.g. Evolutionary Algorithms or Neural Networks), this approach does not consider the how of cognitive information dynamics, but rather the how much and the which. The approach provides models such as for the discovery of sensoritopic maps, the intrinsic generation of goals, the evolution of sensors or the emergence of cooperative collective behaviour.
Lindop Building, University of Hertfordshire
University of Hertfordshire, College Lane, Hatfield, AL10 9AB
Free for members and non-membersLast updated 16th November, 2020 at 4:05pm