My basic idea of perception: send constrained probability distributions bottom-up, collapse them based on hypotheses top-down.
Conversation
Replying to
are you familiar with the "there is no perception" school? Instead you have a self-adjusting system. (sidesteps "who perceives?")