Daniel Polani: "Informational Principles of Embodied Cognition"

In the quest to understand the principles under which cognition emerges in biological organisms and how these principles can be applied to create artificial cognitive systems which espouse these principles, Shannon information has increasingly crystallized as an immensely powerful concept.
Cognition, it turns out, is governed by mathematical laws, invariants and principles which constrain the decision-making of organisms and agents and which can be elegantly expressed in informational terms. This opens a fresh route to analyse, understand and synthesize principles of cognitive systems and their evolution.
The talk will present principles, and applications to understanding the origins and evolution of cognition, consequences for cognitive organization and observations, clarify the role of embodiment in cognitive processing, and finally show some examples how these principles can be used for the generation of autonomous behaviour in agents.