Re: [topicmaps-comment] Re: on the Manhattan project for the KnowledgeSciences

[David Dodds <drdodds42@hotmail.com>]

Hi Rex

I do not have the time to do the work either by the 10th, as you said hope 
somebody in the group can do it.


I hope not to inject a note of pessimism in the goal of our works but as 
Paul Prueitt points out it really isnt the snazzy technology that ultimately 
protects North America, detects problems and anomalies, its PEOPLE who push 
buttons that cause machinery to execute instructions , such as launch. With 
current systems and or Paul's systems in place we are all still at the mercy 
of the military/governmental establishment.

If these advanced systems provide warnings and detections that are then 
consequently sloughed off by one or more humans in the command structure
then even a truely omniscient system such as "DeepThought" would be only of 
passing help to us citizens who rely on the establishment to protect our 
persons and property.

In Cybernetic terms we need to not only adress dramatically the "skill 
level" (lack of it?) in conventional IT and AI (GOFAI), but we also must 
address the (interfering) factors of human factors. IT and GOFAI designs 
tend to ignore or eliminate design aspects of what impact is has having 
HUMANS in the loop. Cybernetic designs INCLUDED them, and so were more 
complete and wholistic. BTW, IT comes from automating bookeeping and sending 
goods or transactions through some conduit, delivery system etc.

Bankers and businessmen, the buyers of IT, dont want any deviation from 
rather simplistic repeated actions, like ledger keeping or shipping records 
etc. Its money! we're dealing with here after all..  ;)

Looking to IT type solutions for any kind of intelligent or adaptive 
operation is a non-sequitor and a nonstarter.

In philosophical circles (What Computers Still Cant Do, which was published 
back in the days when steam powered computers :) ) it has been known for 
YEARS and years that there are philosphically detectable problems with 
rule-based systems (ie Expert Systems), and also, I have personally worked 
with neural net software enough to know the many pitfalls of that area. 
(hint,"nobody is home" in ANNs)

Still, as PPrueitt also said, because the systems which protect us are 
ultimately SOCIAL systems populated by humans (people), who have a great 
deal of power ("discretion") to operate (or not) defensive / offensive 
weaponry at our behest, the *systems analysis* of what needs to be done must 
also be cognizant of the "imperfections" (power tripping, thick headedness, 
need to "control",etc) of the military and government personnel "manning" 
these systems.

Those of us, including myself, who have done computer based modelling of 
human/social systems are already aware that it is "the human factor", the 
imperfections and weaselness of some humans in a human-vetted system that 
make such a system vulnerable and exposed. (of course a system built with 
any of today's technology that controlled weaponry but had no vetting by 
humans in the loop would have us all glowing in the dark wthin a few months 
of the power switch going on.)
When the very powerful radars of the DEW line were first turned on a 
squadron of fighters was sent out to shoot down the moon. Nowadays we have 
nuclear cruise missiles and there ARE pilotless aircarft NOW in the skies, 
computers run them, they are not flown directly by humans. fortunately a 
human or two sits and watches some panels and guages to "oversee" what these 
computer controlled planes are doing.

has anybody seen "Terminator"?

I think that human markup group has a postive part to play in all of this , 
my own HML research shows that HML DAML can be used in conjunction with 
Lakoff-deBono active diagrams to effect some 'understanding' (in a 
"computer") of social behaviours. It looks promising. This sort of thing can 
be used to do modelling of the human aspects of human/machine systems such 
as our defense system.

David Dodds



>From: Rex Brooks <rexb@starbourne.com>
>To: paul <beadmaster@ontologystream.com>, Rex Brooks <rexb@starbourne.com>
>CC: Humanmarkup-Comment <humanmarkup-comment@lists.oasis-open.org>, 
>categoricalAbstraction <categoricalAbstraction@yahoogroups.com>, 
>Topicmaps-Comments <topicmaps-comment@lists.oasis-open.org>, 
>associates@thenation.com
>Subject: [topicmaps-comment] Re: on the Manhattan project for the  
>Knowledge Sciences
>Date: Sun, 02 Jun 2002 20:51:02 -0700
>
>Hi Paul,
>
>I can only speak for myself and say that I just don't have enough
>time for the all-out effort that would be required from me for making
>the June 10, 2002 deadline, not without sacrificing previous
>commitments. However I hope someone in our effort can find the time
>to do so.
>
>The president of American Airlines seems intent on making himself and
>the rest of the airline fleets a stunning example of just how
>shortsighted we can be since he just voiced his opinion that the
>extra security we placed on our air fleets are no longer needed
>because he does not think that the airlines are a target for
>terrorists any longer.
>
>Best of Luck, it looks like we'll need it.
>
>Rex
>
>
>At 9:08 PM -0400 6/2/02, paul wrote:
>>Open public letter..  please forward as you wish.
>>
>>Rex,
>>
>>In reading your message to me, and in reflecting on some parts of other
>>recent conversations, I realize - once again - that it is a hard sell for 
>>me
>>(or anyone) to indicate that there MIGHT be a bias against the biological
>>model of intelligence by computer technologists (and IT marketers) ... 
>>that
>>it is not merely (as you put it) "basic laziness or unwillingness to spend
>>time outside their normal activities and habits" .
>>
>>The deficit to this bias, if it does exist, might be reflected in the
>>limitations that we clearly see in the kinds of artificial intelligence 
>>that
>>folks like the FBI Director think is going to make things a lot better in
>>terms of intelligence vetting.  It is not ranting, on my part, to suggest
>>that AI does not and can not solve the decision support and recognition of
>>novelty problems that are likely to the be next threat that the "system"
>>does not see. Penrose and Rosen and others set the problem out very 
>>clearly.
>>
>>What the FBI "need" for AI will do is put more money, much more money, in
>>the pockets of DoD IT vendors, who think that R&D is a buzz word having
>>something to do with selling half baked product.  As a general principle, 
>>IT
>>vendors are focused on making money, not on solving problems.
>>
>>During the confusion involved in spending this money no one will be able 
>>to
>>criticize the effort based on scientific merit.  And later next year when 
>>we
>>get hit again...  some how the whole claim that AI will make it predictive
>>and preventative will be forgotten.
>>
>>It is as if we (with the help of our news media) can not remember those
>>things that we should remember.  What was the intelligence failure when 
>>all
>>intelligence agencies failed to predict Pakistan starting to test atomic
>>bombs?  It was the notion that - in spite of the clear evidence - no one
>>should expect what actually happened.
>>
>>I would not be making this type of message public except that I am very
>>fearful about the New War taking an unexpected turn that need not occur if
>>openness, truth and honesty were the first principle of our foreign and
>>public policy.  If India attacks Pakistan, and there is a Nuclear war 
>>there,
>>then an American city will suffer a similar fate soon after.  Why?  It is
>>simply because of the tremendous suffering that (predictably) some people 
>>in
>>the East will feel should be shared by those who they feel built the 
>>weapons
>>and made the war a reality.  This is Predictive Analysis Methodology, and
>>this is not understood in the Middle East, and not in our establishment.
>>
>>Society is too powerful to not have real knowledge of what is really going
>>on in the world.  We must find our way to the knowledge age, and in doing
>>so, we must avoid all of the religious fundamentalists and the scientific
>>reductionists also.  We must take the babble that they say openly as how
>>they really feel.  Their intense believes drives them to make their vote
>>count many times more than those who are the mainstream.  War and 
>>confusion
>>is on the path to the rapture, and thus is desirable (to them).
>>
>>What will make human/machine intelligence predictive and preventative is a
>>system of knowledge sharing and the avoidance of making everything useful
>>classified.  This means, oddly enough, that we can win the New War by
>>returning to the democratic practice of informing the public about what is
>>actually reality.  We can lose the New War and the democracy by pumping
>>billons more into IT concepts that are illusions and then so abuse the
>>Constitution so as to invade the privacy of anyone...  and to look the 
>>other
>>way while this invasion of privacy is happening big time by the Microsoft
>>developers community.
>>
>>The common factor is control of social reality by pure blind and greedy
>>economics.
>>
>>My feeling (after hearing the FBI Director make statements that clearly 
>>are
>>not and can not be TRUE in the Sunday talk shows) is a reinforcement that
>>this American administration is going to get milked by IT and the DoD
>>cottage industries (again) over this AI Dream myth.  My feeling reminds me
>>of my feeling about the comments of National Security Advisor Ms. C. Rice,
>>"no one could have expected 9-11" when one frustrated FBI agent had just
>>finished writing an entire book, 3 days before 9-11 took place on the fact
>>that intelligence vetting or the known facts (pre - 9 -11) regarding Bin
>>Ladin's use of US flight schools to train for the next big attack, was 
>>being
>>actively blocked by the administrators of the intelligence systems.  These
>>folks is expensive suits.
>>
>>Clearly there were folks who did expect something, and they did this 
>>without
>>AI and without the full support of the intelligence systems. They opened
>>their eyes and looked.
>>
>>Why is it that we (?) hid information from the public?
>>
>>So the FBI Director wants computers and AI to solve the problem of a
>>Predictive Analysis Methodology?  Well, I  .....  (redacted) ...  he is
>>listening to the same folks that will get a trademark on the term PAM
>>(Predictive Analysis Methodology) and then go and try to find a scientist 
>>to
>>put AI into the cyber warfare systems (without having a clue as to the 
>>real
>>issues involved in the prediction of complex reality).  The Nation's
>>(www.thenation.com) cover asking if the President is clueless is getting
>>close to the mark.  But one hopes that he and our Nation will look closely
>>at this funding the problem and avoiding the solution behavior.  God bless
>>him and our Nation.
>>
>>***
>>
>>The solution is a new kind of computer science, one that recognizes the
>>difference between the cause in a formal system (logical entailment) and 
>>the
>>causes in a natural system (physical entailment) such as a human mind or a
>>human community.  And the solution is a new kind of openness about the
>>failure of not only the human organizations that hid incompetence and
>>laziness, behind the notion of "National Security", but also the 
>>information
>>technology that our society has paid for and paid for and will pay for
>>again... but that has still not addressed the required grounding of 
>>computer
>>theory in some type of natural science.
>>
>>I think I have the argument down to the relationship between reductionism,
>>the defense of a status quo that has perhaps lead us down the wrong path 
>>(to
>>scientific reductionism and AI) and the technical difficulty of the
>>backgrounds that I have learned (over forty years of hard work) ...
>>advanced and pure mathematics and foundations of logic plus a considerable
>>amount of cognitive neuroscience and immunology...
>>
>>Perhaps the single easiest entrance into the grounded understanding of the
>>limitation of the current computer science paradigm is in the computer
>>science community's use (mis-use) of the word complexity.  In the 
>>scholarly
>>work to which I make reference in:
>>
>>http://www.bcngroup.org/area3/pprueitt/book.htm
>>
>>we define complexity in a way that would give a computer an halting
>>condition with formally both a = b and a Not(=) b at the same time.  So in
>>the category theory of Rosen and perhaps Penrose, there simply can not be
>>computational complexity - when defined in this natural science sense.
>>Complexity is a property of all real systems, even any non living system
>>like a computer is complex (has quantum mechanical fluctuation for 
>>example)
>>..... but not a property of a computer program (oddly enough). All 
>>computer
>>processes are simple, even the entire Internet (.. Kat, I say this to you
>>again. .. )
>>
>>The "stratified complexity" view of natural science has huge implications 
>>to
>>computer science and ultimately to the kind of knowledge technology that 
>>the
>>FBI Director needs but does not know about.  Human markup of communication
>>and human behaviors can be made sense of only if one sees what those in 
>>the
>>science of human memory, awareness and anticipation see... that separate
>>levels of organization are responsible and necessary to natural 
>>intelligence
>>and human behavior.
>>
>>Also, using the notion of emergence, one can show that in most cases,
>>natural science would define emergence to be a process that has the 
>>property
>>of being non-reversible - and nothing in the computer's finite state 
>>machine
>>can ever not be reversible.
>>
>>There are questions about the fundamental differences between "simulating
>>the effects of natural law" and having natural law cause motion.
>>
>>The specific argument is given in
>>
>>http://www.bcngroup.org/area3/pprueitt/kmbook/Chapter2.htm
>>
>>
>>So your position is partially correct.
>>
>>But the issue is action perception cycles .. and all that this really
>>entails if we are to develop a "stratified" computational kernel for
>>knowledge technology.  I think that we can do this and will write about it
>>in the next few days.
>>
>>Thank you for looking at "Detection Event in Computational Space"
>>
>>http://www.ontologystream.com/cA/papers/cA-SPS.htm
>>
>>My work on a the computational kernel for knowledge technology is being
>>developed at:
>>
>>http://www.ontologystream.com/admin/KTEcosystem.htm
>>
>>and I now have everyone's agreement that this presentation can be made
>>public.
>>
>>It is only a working concept, and I ask that anyone who is able to bring 
>>the
>>work of the human mark up community to this work to please give me a call.
>>
>>The NIST ATP deadline is June 10th... and I am interested in adding other
>>junior partners to the Knowledge Net Consortium for those who feel that 
>>they
>>should be in involved in this.
>>
>>My work on the proposed Manhattan Project for the Knowledge Sciences is 
>>at:
>>
>>http://www.bcngroup.org/area3/manhattan/sindex.htm
>>
>>
>>Paul Prueitt, PhD
>>703-981-2676
>
>
>--
>
>----------------------------------------------------------------
>To subscribe or unsubscribe from this elist use the subscription
>manager: <http://lists.oasis-open.org/ob/adm.pl>




David Dodds


_________________________________________________________________
Join the world’s largest e-mail service with MSN Hotmail. 
http://www.hotmail.com