From: "Ben Goertzel" <[EMAIL PROTECTED]>
However, in the current day, I would say that we can list some principles
that any successful project must comply. Anyone want to start the list?
Sergio Navega.
Sergio,
While this is an interesting pursuit, I find it it much more difficult
than the already-hard problem of articulating some specific AGI
design.
My view is that mind is largely an emergent quality associated with
certain types of systems. We may describe the properties of mind ...
I have tried to do that in some past writings ... but, then, the set
of underlying systems that may give rise to the properties of mind may
well be a VERY LARGE SET.
Ben,
I agree, it is indeed a very large set. But each property that one
adds to that specification list will likely reduce the number of
possible systems capable of presenting that behavior. I still think
it is imperative to come up with a sufficiently detailed list. But...
The most difficult thing in doing this listing is not to have a
sufficiently small (manageable) set of functional properties, but it
is to avoid the temptation of mixing levels of analysis. When one
speaks of memory, for instance, one has to understand that
this is way too far from conceptual combination or analogical
reasoning (and even more of logical inference). What we need is a
set of sufficiently low (but not lower) level specifications that
can raise some of those high level behaviors. We cannot use the
high level behaviors as "specifications" for our systems, we should
use them only to check if our design is going in the right direction.
So, if there are **general** principles for AGI design, they are going
to have do with the overall emergent behaviors that the AGI design has
to lead to, rather than with specific aspects of the design itself.
At least, that's my feeling at the moment ... I'm of course willing to
be refuted by someone's conceptual breakthrough ;-)
I'm sure I'll not give you that breakthrough (at least for now ;-),
but I understand what you mean. We want those emergent behaviors
(and they certainly are emergent). This is our curse: we want to
design a system using a very large number of possible simple mechanisms
and which only some of them may, when in operation, give us the high
level emergent behaviors we want. This is certainly very difficult, but
it is, as AI history can attest, much better than trying to devise a
system that produces the high level behavior from first principles
(such as logic reasoning or cyc-like KB inferences).
Sergio Navega.
ben
-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/[EMAIL PROTECTED]
--
No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.1.405 / Virus Database: 268.12.8/455 - Release Date: 22/09/2006
-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/[EMAIL PROTECTED]
