[Martin Taylor 920818 18:30]
CSG-L readers might be amused by this paragraph from a mailing on the
Usenet newsgroup "sci.cognitive". I append the author's address in case
someone want to point him here (I don't know the usenet name for the
extension of CSG-L). Note that this is just one paragraph out of a
longish posting. I don't know what HDP (hierarchic dynamic programming)
entails, but from a skim of a couple of earlier messages in this thread,
it seems somewhat similar to the higher levels of PCT. Could be wrong
on that. Here's the paragraph.
ยทยทยท
==========
Dr. Thompson also comments that he doesn't see a role for "goals" in the
AI sense in biological systems. In order to have a clear goals in HDP, you
need to be informed when you have reached certain terminal states. I agree
that biological systems do not have those terminal states, however, that
doesn't prevent them from approximating such states. I suspect there is a
calibration process involved.
--
Harry Erwin
Internet: erwin@trwacs.fp.trw.com
=============
Rick...Don't let the loose canon explode the wrong way, please!
Martin