Richard,
I quickly reviewed your paper, and you will be happy to note that I
had underlined and highlighted it so such skimming was more valuable that it
otherwise would have been.
With regard to "COMPUTATIONAL IRREDUCIBILITY", I guess a lot depends
on definition.
Yes, my vision of a human AGI would be a very complex machine. Yes,
a lot of its outputs could only be made with human level reasonableness
after a very large amount of computation. I know of no shortcuts around the
need to do such complex computation. So it arguably falls in to what you
say Wolfram calls "computational irreducibility."
But the same could be said for any of many types of computations,
such as large matrix equations or Google's map-reduces, which are routinely
performed on supercomputers.
So if that is how you define irreducibility, its not that big a
deal. It just means you have to do a lot of computing to get an answer,
which I have assumed all along for AGI (Remember I am the one pushing for
breaking the small hardware mindset.) But it doesn't mean we don't know how
to do such computing or that we have to do a lot more complexity research,
of the type suggested in your paper, before we can successfully designing
AGIs.
With regard to "GLOBAL-LOCAL DISCONNECT", again it depends what you
mean.
You define it as
"The GLD merely signifies that it might be difficult or
impossible to derive analytic explanations of global regularities that we
observe in the system, given only a knowledge of the local rules that drive
the system. "
I don't know what this means. Even the game of Life referred to in
your paper can be analytically explained. It is just that some of the
things that happen are rather complex and would take a lot of computing to
analyze. So does the global-local disconnect apply to anything where an
explanation requires a lot of analysis? If that is the case than any large
computation, of the type which mankind does and designs every day, would
have a global-local disconnect.
If that is the case, the global-local disconnect is no big deal. We
deal with it every day.
I don't know exactly what you mean by "regularities" in the above
definition, but I think you mean something equivalent to patterns or
meaningful generalizations. In many types of computing commonly done, you
don't know what the regularizes will be without tremendous computing. For
example in principal component analysis, you often don't know what the major
dimensions of a distribution will be until you do a tremendous amount of
computation. Does that mean there is a GLD in that problem? If so, it
doesn't seem to be a big deal. PCA is done all the time, as are all sorts
of other complex matrix computations.
But you have implied multiple times that you think the global-local
disconnect is a big, big deal. You have implied multiple times it presents
a major problem to developing AGI. If I interpret your prior statements
taken in conjunction with your paper correctly, I am guessing your major
thrust is that it will be very difficult to design AGI's where the desired
behavior is to be the result of many casual relations between a vast number
of active elements, because in such system the causality is so non-linear
and complex that we cannot currently properly think and design in terms of
them.
Although this proposition is not obviously true on its face, it is
arguably also not obviously false on its face.
Although it is easy to design system where the systems behavior
would be sufficiently chaotic that such design would be impossible, it seems
likely that it is also possible to design complex system in which the
behavior is not so chaotic or unpredictable. Take the internet. Something
like 10^8 computers talk to each other, and in general it works as designed.
Take IBM's supercomputer BlueGene L, 64K dual core processor computer each
with at least 256MBytes all capable of receiving and passing messages at
4Ghz on each of over 3 dimensions, and capable of performing 100's of
trillions of FLOP/sec. Such a system probably contains at least 10^14
non-linear separately functional elements, and yet it works as designed. If
there is a global-local disconnect in the BlueGene L, which there could be
depending on your definition, it is not a problem for most of the
computation it does.
So why are we to believe, as your paper seems to suggest, that we
have to do some scan of complexity space before we can design AGI systems?
In the AGI I am thinking of one would be able to predict many of the
behaviors of the machine, at least at a general level from local rules,
because the system has been designed to produce certain types of results in
certain types of situations. Of course, because the system is large the
inferencing from each of the many local rules would require a hell of a lot
of computing, so much computing that a human could not in a human lifetime
understand everything it was doing in a relatively short period of time.
But because the system is an machine whose behavior is largely
dominated by sensed experience, and by what behaviors and representations
have proven themselves to be useful in that experience, and because the
system has control mechanism, such as markets and currency control
mechanisms, for modulating the general level of activity and discriminating
against unproductive behaviors and parameter settings -- the chance of more
than a small, and often beneficial, amount of chaotic behavior is greatly
reduced. (But until we actually start running such systems we will not know
for sure.)
It seems to me (perhaps mistakenly) you have been saying, that the
the global-local disconnect is some great dark chasm which has to be
extensively explored before we humans can dare begin to seek to design
complex AGI's.
I have seen no evidence for that. It seems to me that chaotic
behavior is, to a lesser degree, like combinatorial explosion. It is a
problem we should always keep in mind, which limits some of the things we
can do, but which in general we know how to avoid. More knowledge about it
might be helpful, but it is not clear at this point how much it is needed,
and, if it were needed, which particular aspects of it would be needed.
Your paper says
"We sometimes talk of the basic units of knowledge-concepts
or symbols- as if they have little or no internal structure, and as if they
exist at the base level of description of our system. This could be wrong:
we could be looking at the equivalent of the second level in the Life
automaton, therefore seeing nothing more than an approximation of how the
real system works."
I don't see why the atoms (nodes and links) of an AGI cannot be
represented as relatively straight forward digital representations, such as
a struc or object class. A more complex NL-level concept (such as "Iraq" to
use Ben's common example) might involve hundreds of thousands or millions of
such nodes and links, but it seems to me there are ways to deal with such
complexity in a relatively orderly, relatively computationally efficient (by
that I mean scalable, but not computational cheap), manner.
My approach would not involve anything as self-defeating as using a
representation that has such a convoluted non-linear temporal causality as
that in the Game of Life, as you quotation suggests. I have designed my
system to largely avoid the unruliness of complexity whenever possible.
Take Hecht-Neilsen's confabulation. It uses millions of inferences
for each of the multiple words and phrases its selects when it generates an
NL sentense. But unless his papers are dishonest, it does them on an
overall manner that is amazingly orderly, despite the underlying complexity.
Would such computation be "irreducibly complex"? Very arguably by
the Wolfram definition, it would be. Would there be a "global-local
disconnect"? It depends on the definition. The conceptual model of how
the system works is relatively simple, but that actual
inference-by-inference computation would be very difficult for a human to
follow at a detailed level. But what is clear is that such a system was
built without having to first research the global-local disconnect in any
great depth, as your have suggested is necessary.
Similarly, although the computation in a Novamente type AGI
architecture would be much more complex than in Hecht-Neilsen's
confabulation, it would share certain important similarities. And although
the complexity issues in appropriately controlling the inferencing a
human-level Novamente-type machine will be challenging, it is far from clear
that such design will require substantial advances in the understanding of
global-local interconnect.
I am confident that valuable (though far less than human-level)
computation can be done in a Novamente type system with relatively simple
control mechanisms. So I think it is worth designing such Novamente-type
systems and saving the fine tuning of the inference control system until we
have systems to tests such control systems on. And I think it is best to
save whatever study of complexity that may be needed to get such control
systems to operate relatively optimally in a dynamic manner until we
actually have initial such control systems up and running, so that we have a
better idea about what complexity issues we are really dealing with.
I think this make much more sense than spending a lot of time now
exploring the -- it would seem to me -- extremely very large space of
possible global-local disconnects.
Ed Porter
-----Original Message-----
From: Richard Loosemore [mailto:[EMAIL PROTECTED]
Sent: Wednesday, December 05, 2007 10:41 AM
To: agi@v2.listbox.com
Subject: Re: [agi] None of you seem to be able ...
Ed Porter wrote:
RICHARD LOOSEMOORE====> There is a high prima facie *risk* that
intelligence
involves a
significant amount of irreducibility (some of the most crucial
characteristics of a complete intelligence would, in any other system,
cause the behavior to show a global-local disconnect),
ED PORTER=====> Richard, "prima facie" means obvious on its face. The
above
statement and those that followed it below may be obvious to you, but it
is
not obvious to a lot of us, and at least I have not seen (perhaps because
of
my own ignorance, but perhaps not) any evidence that it is obvious.
Apparently Ben also does not find your position to be obvious, and Ben is
no
dummy.
Richard, did you ever just consider that it might be "turtles all the way
down", and by that I mean experiential patterns, such as those that could
be
represented by Novamente atoms (nodes and links) in a gen/comp hierarchy
"all the way down". In such a system each level is quite naturally
derived
from levels below it by learning from experience. There is a lot of
dynamic
activity, but much of it is quite orderly, like that in Hecht-Neilsen's
Confabulation. There is no reason why there has to be a "GLOBAL-LOCAL
DISCONNECT" of the type you envision, i.e., one that is totally impossible
to architect in terms of until one totally explores global-local
disconnect
space (just think how large an exploration space that might be).
So if you have prima facie evidence to support your claim (other than your
paper which I read which does not meet that standard
Ed,
Could you please summarize for me what your understandig is of my claim
for the "prima facie" evidence (that I gave in that paper), and then, if
you would, please explain where you believe the claim goes wrong.
With that level of specificity, we can discuss it.
Many thanks,
Richard Loosemore
), then present it. If
you make me eat my words you will have taught me something sufficiently
valuable that I will relish the experience.
-----
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/?&;
-----
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/?&;
