Enshrining a Neural Expansive Universe of Obligation via Collective Cognition and Calling It Good Citizenship
Excerpt: In the rationalistic view [the old transmission of knowledge/textbook/lecture approach], the world is composed of definite objects, properties and relations, and ‘learning’ is the process whereby an agent forms a mental model of the world that correctly describes these features. Learning in classifier systems is acquiring circumstance-specific behavioral propensities that function together to produce reward. That is, the agent is learning how to act in the world, rather than how to describe it.
No comments:
Post a Comment