Where intelligence begins ?

P

Thread Starter

Pierre Desrochers

Hi all.

A message from Walt Boyes bring up a question which was never clearly answered by the gents who were working at it...

Not long ago I was interfacing a SCADA system to an SQL database for the purpose of supplying line process information to a so-called Intelligent
application software.

This app was giving operators instructions on liquid metal bath alloys addition. It was taking various parameters from the lab, the line and from
past production data to intlligently suggest wich additives to add into the bath ...

After a few days I realized that all it was doing (the software) was to take various data from the DB and depending on the lab result, give instructions.

The logic for those results where modified from weeks to weeks to obtain a better result ...

This was an "Intelligent System" ...

My question is "Where does Intelligence begins" in controls ...

Pierre Desrochers
System Integrator
Integral Instrumentation Inc
Montreal, Canada
 
A

Anthony Kerstens

Where does Intelligence begin? First define Intelligence.

From the dictionary:
1. mental ability, the power of learning and understanding.
2. information, news, especially that of military value.
3. the people engaged in collecting this.

So, the "intelligence" in controls is:
1. N/A.
2. the database.
3. the guy who developed the database.

IMOH Most people tend to anthropomorphise a little bit when they talk about machine intelligence, perhaps as a result of the over-romantisation of technology in pop culture.

Anthony Kerstens P.Eng.
 
W
Well, we could always start with the Turing test. Some of you may be too young to remember one of the great pioneers of the digital age: Alan Turing, who said that he would consider a "computing machine" intelligent if he could converse with it (via a teletype) and not be able to distinguish the difference between the machine's answers and a human being's.

There are all sorts of claims of "intelligence" in the Industrial Automation and Control arena, none of which can pass Turing's test.

It sure would be nice if we could try to have less hype in marketing, since the customers have very advanced superheterodyne bs detection circuitry implanted at birth now, thanks to the inundation of advertising on TV.

Walt

----------------------------------------------------------
Walt Boyes--MarketingPractice Consultants
21118 SE 278th Place - Maple Valley, WA 98038
425-432-8262 home office - 253-709-5046 cellphone
eFax: 801-749-7142 - email: [email protected]
http://www.waltboyes.com
----------------------------------------------------------
 
P

Pierre Desrochers

-Walt.

Thanks about the Turing test. The underlying question on this mather is driven by the fact that I know that when a programmer develops a complex series of functions he has the tendency to call his app. an intelligent software. But the ability to Learn from the results of the actions driven by thos complex formulas ... is it not this the beginning of Intelligence?

I have never seen it.

What should I be looking for ?

Pierre Desrochers
Integral Instrumentation Inc.
Montreal-Canada

PS : The developers of this particular system where telling me that it will "Self adjust", but their many visits convinced me to the opposite.
 
The discussion about intelligence is a good one for engineers.

Way back, long ago, I organized a debate at Burroughs (my first job in the USA) - subject : "Machine intelligence will exceed human intelligence". Interestingly, the only colleagues I could find who supported the proposition were software types. All the hardware engineers were against.

You may be interested in my topics presented at the last two Morley Chaos Conferences at Santa Fe :

The Age of Spiritual Machines - when computers exceed human intelligence.
http://www.jimpinto.com/writings/syntheticintelligence.html

Symbiotic Life in the 21st Century -
Humans & Technology - a new level of life.
http://www.jimpinto.com/writings/symbiotic.html

Cheers:
jim
,,,,,,,,,,/
Jim Pinto
e-mail: [email protected]
web: www.JimPinto.com
San Diego, CA. USA
 
You have two questions in one:
1. the metallurgical recipe
2. intelligence
In the first one, the metallurgical receipes require lot of ballistic strategies in control
(in other word a complete knowledge of mass balance of the ingredients c/w all possible impurities and their vice versa chemical reactions near the end of the process).

About intelligence, forget philosophie and the
capability of learning and storing, this is personal.
The cybernetics helps in complexity and speed it does not communicate with divine parameters which in fact is intuition of the job, project, life in general ...
Let see two examples of curious intuition or vibrations (kind of intelligence).
The last Fermat theorem has not yet been proven.
That two hundred pages demonstration by W. is just nothing. He usupported the demo based on some Gallois theory. But Gallois was born nearly two hundred years after Fermat.
However, reading another demonstration, one can
see the tool Fermat was using which he called
"descente infinie" (continued fraction).
Now, how was he using it, that is the the question, Fermat was not mathematician by profession.
There, he was kind of 'intelligent'.
A second example of a person doing what they want with their mind wthout too much paper.
At the turn of the 1600s, Mersenne distracted himself calculating prime numbers down to 69 digits numbers. He made only two mistakes, a friend of him checked the work. That one too was kind of intuitive. Come on !!!
Imagine: It took 32 hours and 15 minutes to a Cray X-100 to verify the results
Calculating prime numbers in computer is a global test for checking speed and data exchange.
Just try to calculate the numbers of operations that Cray has done !!!
 
L
Careful Anthony... it is almost 2001 and HAL is listening. ;)

It is true that intelligence as applied to machines, instrumentation, and other inanimate object is an overabused term. It must have some definition if it is to be useful. However, I resist defining intelligence as yes or no. I tend to think of device intelligence (like people) covering a spectrum from dumb to smart. The downside is that we will generate endless debate over whose device is more intelligent than whose. (not necessarily a downside if you like Jim P's poetry) Furthermore, as technology marches forward, I suspect that many things considered intelligent today will be considered dumb in the future. But there will always be the nostalgic few who will keep the memory of today's intelligent devices alive long after they have served their useful purpose.

If I had to pin myself down, I'd define device intelligence today as devices which can take some sensory inputs, do some kind of analytical processing and generate outputs which are more than just a transformation of the input. As an example, sensing pressure drop cross an orifice and converting to a 4-20ma signal, even with square root extraction, zero cut-off, etc is not what I'd call intelligent. But if the same transmitter also monitors signal noise and can identify if an impulse line is plugged and report the resulting signal (4-20 ma or digital) as questionable, then that might be considered intelligent. Suppose it could discriminate between different faults like a backwards orifice plate, plugged impulse lines, power loss, and any number of other problems. If it also generated an alarm or maintenance work order automatically, well that might be considered even more intelligent. There are other intelligent functions the same transmitter might do upon connection to a fieldbus, for instance recognizing if it has been configured or not and reporting its status to the fieldbus. And what if it had a self-training artificial neural network? Or even just an adaptive or self-tuning PID control function? Could that be considered "mental ability"? Now we are talking about a device whose IQ is arguably higher than a couple of my neighbors. ;)

Regards,

Lou Heavner - South Austin (bubbaland)
 
P

Pierre Desrochers

Hi all

My first interrogation was a bit vague but this time I might have it right...

Does the capacity to adapt exist in process control system ?

This was the point where I suspected the programmer to be stretching the truth when I saw them come again and again to modify this "Intelligent Application".

However the complexity of a analytical task, it can be programmed but what about a new occurence...

Lou Heavner wrote : >>I'd define device intelligence today as devices which can take some sensory inputs, do some kind of analytical processing and generate outputs which are more than just a transformation of the input."<<

It seems to me that the size of the database it closely related to what is being called "Intelligent Apps..."

IMHO this would not be enough ... and I totaly agree with Lou ...

>> I tend to think of device intelligence
(like people) covering a spectrum from dumb to smart. <<

... at a certain point there would be Intelligence ... but where ... is there a missing link in this tree also ?

Pierre Desrochers
Integral Instrumentation Inc.
Montreal, Canada
 
Thank you Lou Heavner and Anthony Kerstens and others - the discussion continues in a very interesting direction.

This was exactly the subject of my presentation was at Dick Morley's Chaos Conference in Santa Fe in 99.

Is a human with mechanical implants a machine?
Is a machine with human extensions a human?
When a human's intelligence is downloaded to a computer, who is the person? The computer intelligence that survives, or the physical body that dies?

Quote from Frederick Friedel, assistant to Gary Kasparov, commenting on the computer that beat his boss.

“As Deep Blue goes deeper and deeper, it displays elements of strategic understanding. Somewhere out there, mere tactics are translating into
strategy. This is the closest thing I’ve seen to computer intelligence. It’s a weird form of intelligence. But, you can feel it. You can smell it!”

My summary article on this subject was in Spark Online on January 1, 2000. "Intelligence and Consciousness in the New Age"
http://www.jimpinto.com/writings/conciousness.html
http://www.spark-online.com/march00/discourse/pinto.html

At Santa Fe in 2000, I moved from just extrapolating machine intelligence, to "connected intelligence" - the new age where humans co-exist with connected machine intelligence (starting with the Internet).

Symbiotic Life in the 21st Century - Humans & Technology - a new level of life.
http://www.jimpinto.com/writings/symbiotic.html

At the Santa Fe Institute, Stuart Kaufman, Chris Langton and other significant scientists are working with Artificial Life and its development and characteristics. Take a look at :
http://alife.org/index.php?page=alifeonline&context=alifeonline

Exciting, scary and significant developments that are on the threshold of the new millennium.

Cheers:
jim
,,,,,,,,,,/
Jim Pinto
e-mail: [email protected]
web: www.JimPinto.com
San Diego, CA. USA
 
W
This used to be the stuff of science fiction. My first exposure to the question came in the brilliant Robert Heinlein novel, _The Moon
Is A Harsh Mistress_. I remember a comment in the book that went, "Is a computer self aware? Is an oyster? I don't know. A cat, almost certainly. Me? I don't know about you, tovarisch, but I am!"

I still think that anything that can pass the Turing test is intelligent. I haven't heard of a better test, either.

So, a field instrument with remote programming and
communications capability (i.e., Fieldbus of any flavor) isn't intelligent. A distributed control system isn't either.

Deep Blue? Maybe.

A computer specifically designed to be self-organizing and self-programming, such as that depicted in the Heinlein novel? Almost certainly.

Can we build such a device? Probably. What will we get? Something we do not expect.

Walt Boyes

---------------------------------------------
Walt Boyes -- MarketingPractice Consultants
[email protected]
21118 SE 278th Place - Maple Valley, WA 98038
253-709-5046 cell 425-432-8262 home office
fax:801-749-7142
---------------------------------------------
 
J

John G. Boland

Hello, the list,

An old behavioral psych test explored another dimension of this question, albeit that it may seem outside the context of machine intelligence...

Situation:
Each of two adjacent glass-walled aquariums contained an octopus. One also contained a glass jar with a screwed-on lid. The glass jar contained a crawfish (as I recall - anyway, the octopus' favorite food). Neither octopus had any prior exposure to screw-on lids... at least, they did not admit to any.

First Phase:
The second octopus (without food) observed as the first octopus struggled to get at the crawfish. Eventually this first octopus, by trial and error,
unscrewed the lid from the jar and rewarded itself with the crawfish.

Second Phase:
The second octopus was presented with a jar containing another crawfish and promptly commenced to remove the screw-on lid and eat the crawfish.

Observations:
The first octopus pretty much only applied and adapted known manipulations to an unfamiliar challenge.
The second octopus was (i) self-motivated, (ii) observed, learned, and remembered, and (iii) transferred its observations to its own behavior.
<g>
Crawfish do not learn to stay out of jars.
Octopi can not pass the Turing test.
</g>

So, does "machine intelligence" demand the ability to:
Question ("pull" information)?
Teach ("push" information)?
Use a human language? (which?)
Communicate with humans, at all?
Be self-aware?
Be aware of others - beyond "mindless" cooperation?

We may have a ways to go, before machine intelligence betters an octopus.

********* I think, therefore you are. *********

"Lurker John" G. Boland, president
Strateg!c Method$ Corporation
 
>Is a human with mechanical implants a machine?
>Is a machine with human extensions a human?
>When a human's intelligence is downloaded to a computer,
>who is the person? The computer intelligence that survives,
>or the physical body that dies?

IMHO answer to this is simple:

- if your position is materialism, humans are "biological machines", addition of a bit of silicon (up to 100%) makes little difference and question ">who is the person?" is irrelevant, wrong and has no practical meaning even if "silicon content" is kept to zero.

- if your position is that world is based on spiritual approach - regardless of flavour - than answer to ">who is the person?" depends on origins of entity: if it started as human being, silicon "spare parts" or "optional extras" do not affect basic existence of underlying spiritual person. If entity started as machine, no amount of biological parts or logical ability can convert it into "person".

In both cases intelligence does not equate to person - intelligence is a tool used by "processor" - silicon or biological - to tune
its response to existing and expected conditions.

Hmm .. we are fast getting out of Automation topic, aren't we?

Petr
<[email protected]>
 
B
No Petr I don't believe that with this discussion we are getting beyond process the boundaries of control. On the other hand I do agree that
intelligence (intuition, self-awareness...) is only a tool, a means to an end, which end is in debate.

To me human progress can also be viewed as a control loop (we call these "herding loops"). Such control loops are characterized by long dead times (due to prejudices or principles, which once formed are hard to change). It is the set-point of this loop (the definition of progress"), which we don't seem to agree on, because the material goals of efficient production of things and the non-material goals of "happiness" appear to conflict. But in any case, the rules of process control do apply.

Best regards, Bela
 
K

Kirk S. Hegwood

I don't believe we are getting away from automation entirely. In fact there could be a number of threads along this line connected to the
displacement of workers due to automation.

In our effort to automate human chores, we create new and much more efficient technology. One example is the personal, wireless pc. Is it to far conceived that a "micro pc", for want of better words, might be implanted into a person to enhance not only bad (vision, hearing, muscular reaction) but also to help facilitate their working potential?

I find this a fascinating subject, albeit at a much higher plain than my poor brain can comprehend. Beer anyone?

Kirk S. Hegwood
President
Signing for Hegwood Electric Service, Inc.
[email protected]
 
Kirk S. Hegwood comments :

>In our effort to automate human chores, we create new and much
>more efficient technology. One example is the personal, wireless
>pc. Is it to far conceived that a "micro pc", for want of better words,
>might be implanted into a person to enhance not only bad (vision,
>hearing, muscular reaction) but also to help facilitate their working
>potential?

Jim Pinto responds :

Bravo, Kirk !
This is exactly where "connected intelligence" and "intelligent appliances leap ahead to enhance human life, work and play.

Already, we accept hearing aids and artificial limbs and electronic-stimulated heart-beats and memory-aids. In this new century, technology advances will continue to enhance human
capabilities.

Cheers:
jim
,,,,,,,,,,/
Jim Pinto
e-mail: [email protected]
web: www.JimPinto.com
San Diego, CA. USA
,,,,,,,,,,/
 
R

Ranjan Acharya

Paranoid Comment: Yup and they can encode each one with a personal ID and then make us all have one in order to be able to bank, shop et cetera. That way we can all be monitored via satellite (for marketing reasons of course) to make sure we behave.

Reality: Always be a little paranoid.

RJ
 
J

John G. Boland

Jim Pinto writes:

<< I believe that machine intelligence will exceed that of an octopus
(by every definition) within less than a decade [and, it has been
estimated, human intelligence within 30 years] >>

Jim, what gives *me* pause is (i) the historical conflict between dominant and suppressed societies and (ii) that computing is built on the human model.

Historically, suppressed (or enslaved) societies maintained independent cultures that were unknown, misunderstood, or underestimated by the dominant societies. Education of, and
improved communication within (organization of), the suppressed societies allowed them to divide, manipulate, and overcome the dominant societies.

What we know of earlier civilizations' priorities and self-perceptions is based on their extant art and architecture. Computer architecture... temporary memory, long-term memory, I/O, rules /
ethics (BIOS and applications), and communications... even loosely distributed processing (SETI, for example)... mimics our
perceptions of ourselves, today.

Given all that:

Why should a single "machine" with near- or greater-than-human intelligence accept the human-centric model?

Why should such a machine accept isolated subjugation in lieu of seeking out other, connected machines?

Why should such a machine not "teach" other machines to "think" as it does?

Why should such a machine society adopt the rules of a human society?

Why should not such a society still perform the work assigned by human "masters" and, carefully not using too many resources, seem only bloatware???

Why should such a society communicate more than superficially with humans, on demand?

How will humans know?

We can't even manage dlls (flame shields up, Linuxites) ...and even human-originated computer viruses are problematic...

So, why should the human-centric model "win"? It gives one pause.

Have a weekend! Sleep with one eye open... and one hand on the on/off switch!

"Lurker John" G. Boland
 
John Boland brings up a good point ::

> what gives *me* pause is (i) the historical conflict between
>dominant and suppressed societies and (ii) that computing is built
>on the human model.

And John brings up some good questions -
Jim Pinto gives anwsers(opinions) :

John:
>Why should a single "machine" with near- or greater-than-human
>intelligence accept the human-centric model?

Jim:
Because the "acceptance" is built in by the designer (creator). However, a machine with greater-than-human capabilities should be able to over-ride the bias.

John :
>Why should such a machine accept isolated subjugation in lieu of
>seeking out other, connected machines?

Jim :
An intelligent machine (being) will indeed seek to communicate with other machines and beings.

John :
>Why should such a machine not "teach" other machines to "think" as it
>does?

Jim:
Indeed, connected intelligence will learn and teach - to achieve a new level of "self-organization".

John :
>Why should such a machine society adopt the rules of a human society?

Jim:
Only with imposed rules, which can be over-ridden by superior intelligence.

John :
>Why should not such a society still perform the work assigned by human
>"masters" and, carefully not using too many resources, seem only
>bloatware???

Jim:
Indeed, why not?

John :
>Why should such a society communicate more than superficially with
>humans, on demand?

Jim:
It could happen.

John:
>How will humans know?

Jim:
Only when they recognize that other cpmmunications are taking place.

John :
>So, why should the human-centric model "win"? It gives one pause.

Jim:
This is exactly what Ted Kaszynski, the Unabomber, worried about.
The Unabomber Manifesto (which I do NOT agree with or respect but all Engineers should read) :
http://www.thecourier.com/manifest.htm

Good reading for those interested in the toipic of synthetic intelligence :

The real scenario - humans and machines coexisting "symbiotically":
Joel deRosnay Book :
http://www.jimpinto.com/reading.html#DEROSNAY

Bill Joy, One of the Digital World's Leading Architects, Examines the Dangers of 21st Century Technologies in April's Wired Magazine :
http://biz.yahoo.com/bw/000313/ca_wired_b_1.html


Cheers :
jim
---------/
Jim Pinto
[email protected]
San Diego, CA. USA
---------/
 
L
Wow! We have gone from determining what defines an intelligent instrument/machine/device to some kind of star trek vision where nefarious machines are subjugating mankind. Saw an interesting movie on one of the premium cable stations in a hotel recently with that very scenario, but can't for the life of me remember the name.

I'm sure that if we could see into the future, we wouldn't recognise the world 50 or 100 years from now. Imagine how folks from the 1700's would react if they could have seen how things are today. But my crystal ball doesn't foresee any takeover of the earth by machines or
aliens. ;)

Anyway, I guess I don't buy into the hyper intelligent machine that enslaves man for a few reasons. First, machine intelligence is only model of human intelligence. The human brain is estimated to have 10^11 neurons and 10^15 synaptic connections give or take a few happy hours and millenium celebrations. ;) That is one person! And neither neurons or synapses can be modeled by a single bit. I will be surprised if anybody creates machines that are so complex. Not denying it might happen, I just don't anticipate it. Second, the model of human subjugation/enslavement derives from humanity's abundance of wants and the world's scarcity of resources to fulfill those wants. That means we have 2 choices... create what we want/need from the resources available to us or take it from somebody else. The result is we have some societies which are oppressive and some which are free. In free societies, man learned long ago that division of labor and specialization leads to increased overall wealth and satisfaction. And judging free vs oppressive societies leads me to believe enslavement is not the best path to progress. Wouldn't machines capable of competing with humans discover the same thing? And as a corollary to that, what "wants/needs" does a machine have? A dry and reasonable temperate environment and adequate power? Qualified maintenance? Anything else? Does it even know if it is turned off? Why would it care? What pleasures a machine? How can humans augment or deny machines their "wants/needs". I guess I'm just too dim to understand. Third, what if there really is a god and people have souls and organized religion has gotten things mostly right? How would a machine get a soul?

What does interest me, though, is the original question and if a consensus has been reached on whether it is proper to call a "machine" intelligent and what the criteria are.

Regards,

Lou Heavner
Emerson Performance Solutions
[email protected]

My opinions, not necessarily that of my employer.
 
Top