[From Oded Maler 920812]
(Bill Powers, yesterday):
What would happen to AI and AL models if all computations
were limited to 1 percent accuracy? ... etc.
The point is that there is some level where we (seem, at least, to)
work in discrete terms, e.g., "if my PC doesn't boot then I replace
the diskette" ,etc. The underlying assumption is that there are
lower-level modules, employing feed-back all the way down, making
sure that the abstract action "replace the diskette" is being performed
in a noisy world (in fact this is a too complicated example because
even sub-tasks of this action are not yet explainable *in detail* in
PCT style - try running Little Man in a non-uniform environment with
obstacles, non-reversible consequences, dead-ends [and this is by
no means an attempt to dismiss its achievements - just to note that
there are many hard problems in the highr-levels]).
Now, it's not clear what 1% accuracy means in this abstract world,
because the disk-drive cannot be 99% pregnant - and if I buy from you
a low-level black box, I trust that it will work. If you mean
inaccuracy in the higher-level, that is, replacing the diskette cannot
be performed or it may not solve the booting problem, and maybe
kicking the table will solve it, then you come to the realm of the AI
methods - searching among alternatives for actions, an activity that
needs a lot of creative imagination (more than I have for the moment)
in order to be perceived as gradient following, as servoing for
complex perceptions etc. The people who investigate such phenomena
surely abstract away the lower levels in the same way you abstract
away the micro-structure of neural and muscular cells.
Now I come back to my basic rhetoric question: goal-achieving in this
abstract world with simplifying assumptions should be much easier than
in reality - the lower-level black boxes are assumed to be perfect,
there is no conflict in actions, and still it is very hard to to give
tractable way of solving everyday problems. Why? What is wrong in our
verbal account of our introspection concerning the way we operate in
this level? Do the higher-level percepts and references carry with
them some concrete features of the lower level ones from which
they are constructed/grounded in such a way that the solution "flow"
instead of being searched-for/calculated?
I don't expect immediate answers..
Best regards
--Oded