Text seems brittle because it was tried and it did not work. But neither did
visual, robotic, or other sensor-based AGI. If the brittleness criticism was
based on a lack of substantial achievement in spite of the effort, then the
brittleness criticism would have to be applied to all AGI modalities. Of
course knowledge that is gathered only through text is going to be brittle in
the sense that it would not be able to achieve the range of understanding that
human beings can achieve, but the use of cell phones or robotics are not going
to create genuine human experiences either.
The only conclusion, based on the acceptance of a general lack of substantial
advancement in the field, is that we do not have basic AGI because computers
cannot achieve general intelligence or general intelligence needs even more
advanced hardware than we have or there has been something important missing in
AGI research.
Something that Bayesian enthusiasts never talk about in these discussion groups
is how can a mostly independent learning system make the distinction between
those kinds of situations where Bayesian methods can be used to combine
different sources of data from those cases where different sources of weighted
values can't be combined or have to be combined in a certain way. The lack of
discussion on this subject, which seems like a central issue to me, is
indicative of a major gap in the basic theories of Bayesian AGI. Of course a
few people get this but so what. If they have it figured out then they should
be able to demonstrate it in an actual feasibility test, or at least explain it
to their fellow Bayesians.
The lack of appreciation for simple AGI feasibility tests, especially since it
is based on a belief system that you have to have it all figured out for it to
be AGI, also shows that there is something missing in our basic AGI theories.
I agree that the dismal history of the pronouncements that some interesting new
theory in AI explained everything is pretty discouraging but that does not mean
that there are not some simple theories that could be used as a basis for AGI
and that there is not a simple process that could build AGI knowledge.
Jim Bromer
Date: Sat, 20 Jul 2013 01:26:50 -0500
From: [email protected]
To: [email protected]
Subject: Re: [agi] A Very Simple AGI Project
On 7/16/2013 4:25 PM, Jim Bromer wrote:
I would like to create an initial feasibility test,
using a text-based IO, that would show the potential for
intelligence across a broad range of subject matters (within
that IO modality.) I am not worrying about writing something
that would be scalable to adult human level AGI.
I think text is a brittle representation of an elastic reality.
I believe that there has been something missing in
AI/AGI. Someone needs to show how one might create a good
base for intelligently acquiring knowledge (ie using both
rational and creative methods) which might be scaled up with
some future computers system. The sense that narrow AI can be
pushed beyond human capabilities in certain human games that
once seemed to demand higher general reasoning is a little
unexpected and hard to understand without concluding that
there must be some very basic AGI ideas that haven't been
discovered.
I think learning is about observing and recording scripts (sequences
of observations and reactive actions) -> as memories and habits.
Then we observe our context and follow a suitable script, yet we are
free to change mid-script to another script if we observe that the
context has changed. We can do this internally and call it planning.
But we should still be free to change the planned script during
real-world action.
Possibly, there is a Bayesian selection of appropriate script.
Play involves the exploration of possible scripts; sometimes we just
throw our blocks in the air to see what happens.
Anyway...
Lately I've been thinking that putting AI on cellphones and tablets
could assist the problem of exploring the real world and embodiment
a bit... through the camera(s), gyroscope, microphone, and GPS. I'll
wait for at least the third generation of google glasses and it's
competitors before I endorse that... but that technology could be
similarly helpful.
And those are my thoughts for the month....
-- Dimitry
AGI | Archives
| Modify
Your Subscription
____________________________________________________________
New Policy in Mississippi
March 2012-Drivers w/ no DUIs may qualify for $9/week car insurance...
TheFinanceAuthority.com
-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription:
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com
