[Avery Andrews 950623]
(Hans Blom & others, various posting)
I haven't been able to follow the model-based control discussion too closely,
but I certainly don't grasp why there is an argument: a `world-model'
*is* a kind of perception, even if rather heavily processed; surely the
only significant questions are (a) what kind of processing are required
to get particular jobs done (which Hans would surely know a lot more
about in his field, than we would) (b) what kinds are in fact employed
by living systems. And there's no sense in being doctrinaire about this
either. If bees can use integrators to control modelled distance
from a food source, surely we can managed some reasonably sophisticated
stuff too.
On another topic, alongside of reorganization and `reinforcement', there's
the problem of one-shot learning. Earlier in the week I learned that
the standard capture-release-and-recapture method of monitoring the
movements of lizards doesn't work for the striped legless lizard
Delma Impar (a local threatened species with a penchant for living in
places where people want to put shopping centers), presumably because
they have some kind of perceptual or cognitive ability such that when they
fall into a standard pitfall trap once, they learn something that
enables them not to fall into one again. There's no reason to think that
they're reptilian rocket scientists, so presumably it's perceptual, but
it still seems to work pretty reliably from a single experience.
Avery.Andrews@anu.edu.au