(C)1995 Lee Kent Hempfling All Rights Reserved
KEYWORDS: Intelligence, Artificial Intelligence, AI, NN, Cased Based, Object Based, Brain, Thinking, Cognition,
Consciousness.
ABSTRACT:
We begin with a dichotomy. Intelligence and Artificial Intelligence. They are not one in the same. Due to the so far
inability of any recognizable authority on the topic there has not been a sufficient definition of what Intelligence actually
is. Yet there is an intent search to create Artificial Intelligence. The difference between the two is enormous. In this paper
I will argue the definition of Intelligence and the definition of Artificial Intelligence and the chasm of intellect that
separates the two.
1. WHAT IT IS NOT
1.1 When man first began the search for understanding of the brain there were forests and trees and sand and water.
There was nothing different from today except the knowledge of the world that is better understood and better applied.
The intelligence that first questioned the intelligence is no different now than it was then.
1.2 So is intelligence then to be considered to be knowledge? Knowledge is the result of intelligence and the current
research and application of data base systems and knowledge bases is a reverse evolution which will never digress to the
utilization of it until such time that a viable definition of Intelligence is forthcoming.
1.3 According to The Artificial Intelligence Dictionary, (Ellen Thro) Artificial Intelligence is a "Multi-disciplinary field
encompassing computer science, neuroscience, philosophy, psychology, robotics, and linguistics; and devoted to the
reproduction of the methods or results of human reasoning and brain activity."
1.4 So the search is on and concentrated in the above fields each vying for posture and seeking the stake to the claim of
having found the solution to a question no one has adequately defined.
1.5 By the literal use of the word "Artificial" the multi-disciplinary field has committed itself to acquire something not in
the least bit real. It has succeeded at that.
2. THE METHODS
2.1 Perhaps the most prevalent system for replication of human reasoning is the Expert System. An application of
IF-THEN-ELSE rules to problem solving and classification. Claimed as including decision making and configuration it is a
wonder such decision process has not generated a definable meaning of Intelligence. Perhaps it is because there is none
in an Expert System.
2.2 If there was an instruction set of IF-THEN-WHY-THEN there could stand the chance of intelligence in such a system
but in order to ask the question, "WHY?," the system would have to be aware it contained information that made sense
not only in and to itself but in a grander scheme of things as a part of a whole of information. With far diverse concepts
being comparable.
2.3 There is no algorithm published or in use today that affords the WHY. Now it can be said that if enough IF-THEN-ELSE
statements are linked together a semblance of WHY will emerge but that then brings to mind at what point does the
instruction either double back on itself or come to a halt? Because if it comes to a halt there is a limit on asking WHY and
if it doubles back the same WHY will be asked which means there was no comprehension of the answer to it in the first
place.
2.4 Resolution of problems without actually solving anything is afforded by Case Based Reasoning. Filters applied to a set
of circumstances drawn from memory are whittled down to what worked before (given that what happened has
happened before.) But is there intelligence in such a system? No. It is a reverse of the IF-THEN-ELSE based instruction in
cases or concepts and not in minute specificity until the limit of the IF-THEN-ELSE rules are reached and then the answer
is brought forth. It may not be the answer. It may be just the least common denominator. But it will be the result of the
system. Not that it matters. The system working such as chopping down a tree a piece of the trunk at a time from the
bottom until reaching the applicable leaves at the top.
2.5 Constraint Programming is the opposite of Cased Based Reasoning as it starts by pruning away non applicable data
to result in a limiting of the amount of processing required to arrive at an answer. It works fine in budget minded tasks to
reduce a tree from the top down to its base roots only to find that a tree is not worth anything if it is only roots. Yet it is
used for financial applications and others where the answer is not necessary to be specific just general.
2.6 Neural Networks are the optimum absurdity in the search for an intelligence. A NN spawns an internal matrix of
weights which represent a non-linear relationship between the variable in a specific problem (Smart 1994). Once the NN
has optimally learned by using data base records with a known outcome, it is presented with a set of records with an
unknown outcome. (Smart 1994) This allows testing of how well the NN has learned to differentiate between cases with
different outcomes. (Smart 1994) The learning cycle is repeated until the system has reached a desired level of
performance, or when patterns from new historical records need to be learned and recognized. (Smart 1994) But is a NN
intelligent? No. It is like filling a large upside down cone with a hole at the bottom with a large amount of white marbles
and one red marble with white marbles falling out the bottom until the red marble falls out. Where the red marble fell out
is then considered to be the point of learning and the processes is repeated adinfinitum until the process is duplicated
and that is defined as a replica of the neural process of the brain. How ridiculous. Or in the opposite it can be said that
the NN grows the tree to learn that it grows only to grow another one until it reaches a point where it knows how a tree
grows and is then presented with a tree that doesn't grow as fast or as tall and continues this until it determines it not to
be a tree. It works just fine in applications where data is hard and conclusions are YES or NO such as fraud detection and
credit applications. But it is not intelligent.
2.7 Genetic Algorithms inspired by Darwinian evolution assumes that a large potential of solutions is present and whittles
down the solutions as they each strive for survival with the surviving solution determined to be the optimum solution. Try
telling that to the turtle. It can neither run nor adequately defend itself yet it has survived as long or longer than most
highly survivable creatures. When good rules of classification are not known this system works like throwing the
collection of marbles on a billiard table and shaking it until the heaviest and least round marble remains. It is declared the
winning marble when in fact it survived because of its imperfections. As offsprings are born in mutated chromosomes the
verbiage is beautiful but the process is not intelligent. Not at all.
2.8 Knowledge has been considered to be intelligence. Stuff a machine with knowledge and it will know. No it will not.
It will be a machine stuffed with knowledge.
3. SO WHAT IS IT?
3.1 From the observations above of the methods of Artificial Intelligence it can be seen that in fact the methods used are
not intelligent and are in essence replications of the result of intelligence each approaching a different aspect of the
result of intelligence to apply that aspect to different cases and different situations.
3.2 What that is, is Object Based Reasoning. (Hempfling, 1995C)The entire study of replicating human reasoning is not
based in what reasoning is it is based in what reasoning accomplishes. So therefore the functions performed by all AI
systems is nothing more than the cover doing the job when it is what is inside that actually accomplishes it.
3.3 By replicating the accomplishments it is easy to understand the methodology used. After all if I wanted to write a
Sheakspear novella I would write a Sheakspear novella. I would write it and I would see that it sounded like a Sheakspear
novella, it looked like a Sheakspear novella and it smelled like a Sheakspear novella (how ever one smells) so it must be
a Sheakspear novella. But did Sheakspear write it? Did I use his methods or did I replicate his results? Trivial questions. I
was not trying to replicate Sheakspear I was trying to replicate his novella.
3.4 The brain is our Sheakspear. It performs functions and those functions result in what we observe and we replicate the
observances and call them functions. The observances or outcomes of the brain are not intelligence. The process that
arrives at those outcomes is.
4. A DIVERSION
4.1 Let us divert the issue. Isn't that what has happened in modern science? Not only do we not have any idea of what
intelligence actually is we have changed the subject. When the going gets tough the tough change paths. Or something
similar to that.
4.2. While we have as yet not arrived at a fundamental paradigm of intelligence we have shifted gears and are becoming
engrossed in consciousness. After all, wouldn't the most blatant of results be the Object if our intention was to replicate
the results? Of course it would.
4.3 From Can Physics Provide a Theory of Consciousness? A Review of Shadows of the Mind by Roger Penrose ( Baars
1995):
Central to SOTM is Penrose's contention that contemporary science has failed
to understand consciousness. There is more than a little truth to that --- if we exclude
the last decade --- but it is based on a great historical misunderstanding: It assumes
that psychologists and biologists have tried to understand human experience with
anything like the persistence and talent routinely devoted to memory, language, and
perception. The plain fact is that we have not treated the issue seriously until very
recently. It may be difficult for physicists to understand this --- current physics does not
seem to be intimidated by anything --- but the subject of conscious experience, the
great core question of traditional philosophy, has simply been taboo in psychology and
biology for most of this century. I agree with John Searle that this is a scandalous fact,
which should be a great source of embarrassment to us in cognitive psychology and
neuroscience. But no one familiar with the field could doubt it.
4.4 Routinely devoted to memory, language and perception. Excuse me here but where is the reference to intelligence?
The issue that has not been treated seriously is the issue of Intelligence. Why is that? Perhaps it is because so many
protocols, as listed previously, have defined Intelligence by the very nature of their making outcomes and since
outcomes is the point then it is time to foray into consciousness.
4.5 But I ask the reader: Q: Where does all of this come from? A: It comes from the brain. The brain takes in data and
spits out data we observe as things we then try to copy. Consciousness is a result just like language is a result. Perception
is the act of input and memory is the storage of previous input and previous Intelligence. Just like there are not numerous
brains contributing to all of the outcomes we perceive there are not numerous methods of brain processing. There is only
one method and that method results in the things we spend our research money and our lives trying to understand when
the world as a whole has not a single clue as to what that one underlying process is. Yet we continue the outcome,
Object Based Reasoning (Hempfling, C 1995) approach to replicating human reasoning. Blindly going where all men
have been before. On the beaten path. Off the subject.
4.6 Reference: [STAR] Penrose is Wrong (McDermott, 1995) A Review Of Shadows Of The Mind.
A clue might be this sentence on p. 373: "It is only the arrogance of our
present age that leads so many to believe that we now know all the basic
principles that can underlie all the subtleties of biological action." Penrose wants
to do battle against the arrogance he perceives, especially in the AI community, regarding the problem of consciousness. It is true that AI has, from its inception,
had the ambition to explain everything about the mind, including consciousness. But is this arrogance? Or merely the sincere adoption of a
working hypothesis? If someone wants to work on the problem of mind, it
seems to me that he must choose among three options: treat the brain as a computer, and study which parts compute what; study neurons, on the
assumption that they might be doing something noncomputational; or work in
a seemingly unrelated field, like physics, on the off chance that something
relevant will turn up. In any case, no matter which tack is taken, one gets
mighty few occasions to feel arrogant about one's success. Neuroscience and
AI have made definite progress, and so has physics, for that matter, but their
successes haven't resulted in a general theory of mind. If anything, AI seemed
closer to such a theory thirty years ago than it seems now.
4.7 It sure did. Back then the question was Intelligence. Now the question is Consciousness without so much as a single
real published assertion as to the complete picture. Arrogance is after all only the human trait of defending one's
ignorance.
4.8 From: Large Scale Analysis of Neural Structures (Merkle, 1989)
Even after decades of research we still seem no closer to duplicating the
"common-sense" intelligence that is the holy grail of AI (artificial intelligence)
research. While many major advances have been made -- some immensely
valuable -- they work by avoiding the central problems of AI rather than by
solving them. While we might discover how to duplicate natural intelligence on
a computer tomorrow (after some truly brilliant breakthrough) yet we might
also find the solution eluding our grasp for years, decades, or longer. This
uncertainty is widespread in the AI community -- to quote Marvin Minsky "...the
process could take 500 years ... or it could be just around the corner!" (Minsky,
1985).
Despite the uncertainty about when, eventual success seems assured. The
fundamental reason for long-term optimism in AI research is the existence of
natural intelligence -- both in lower animals and in humans. If nature can do it,
we can do it too -- eventually. Unfortunately, it is unclear how long such
success will take. If evolution took 100 million years, how long will we take?
Significantly less -- but even one thousand years (about one hundred thousand
times faster than nature) is still a very long time.
Direct analysis of natural systems might well be faster than rediscovering the
basic principles of intelligence from scratch. Neurological research has already
had significant influence on vision research (largely because the retina and visual
areas are among the best studied neural systems.) It has also influenced work
on motor control and has inspired the "neural network" approach to AI which
has recently aroused a great deal of interest and has already led to a commercial
product. A better understanding of natural systems will increase this influence
-- a deep understanding of natural systems might make this influence a
dominant factor in the design of AI systems.
4.9 Yes. We have returned to the reality of the topic at hand. Physics is the subject of other papers so I shall not delve
into it here. Consciousness is a result of the one process of the brain and that we shall delve into.
5. WHAT IS IT? - REVISITED
5.1 It is not a new idea. It is no breakthrough in the understanding of intelligence to state that for once Aristotle was
right. Comparison. Why do you think the scientific method employs so many drawn comparisons for explanation of
otherwise non comparing events and concepts?
5.2 It is the method of the brain. From the input receptor to the muscle. From the microscopic to the macroscopic.
Comparison rules the roost of mental activity. But comparison alone results in more than what comparison is used for as
an outcome. When one person describes a visual experience by relating it to something the receiver of the tale has
knowledge of that is considered a form of comparison. But it is not the comparison used in the brain.
5.3 The comparison used in the brain is a computational comparison. That means that all data must reside in the same
form in order for one event to be compared to another and to have a result of that comparison to be used again, as in
memory recall.
5.4 This is where the various methods of implementation of Artificial Intelligence fail. They perform various and non
comparative functions. Such is the case when outcomes are used to based a paradigm. The results of the brain are varied
and dedicated by their respective physical limitations. An arm can not see. So an arm structure would not be used to
perform a visual function. But the arm supplies the brain in the same form of electrochemical messages with feedback of
the arm's motion and condition. An eye can not hear yet the brain performs a function of processing on both inputs in
the same form of device. The neuron. In the same form of connections. This is possible because the brain functions in a
single process for all inputs and outputs. That process is then comparative.
5.5 The brain's process is an establishment of values of input data. Each various type of input (determined by the form of
input receptor) must feed a single form of processing which makes the inputs regardless of the function they carry out
(vision, hearing, taste, smell, pressure, temperature, balance, etc.,) required to send their signals in a harmonious and
equal form. That form is the subject of the beginning of understanding Intelligence.( Hempfling, B1995)
5.6 In an attempt to understand the brain many take the position that the brain is indeed a very low power of
computation. After becoming so accustomed to the ‘power' of faster and faster digital computers the feeling is that
digital speeds are not in the brain so therefore the brain must not be very smart. After all the brain's neurons are so slow.
They can't possibly store enough information to account for human thinking. So it is with the defense of one's ignorance.
5.7 Remember the turtle? It survived against great odds for millions of years. It is so slow. Why has it survived? It is
efficiency. Speed is not efficient. It is required when making a sine wave out of a square wave. That is what digital
computing does. The more sine in appearance the faster the square wave need be. The more natural the display
(Hempfling 1995). The brain is such that it is so efficient its power of computation is performed at speeds so slow they
appear to be dormant when compared with MHZ and GHz speeds in digital processing. Yet the comparison is not
possible. It is not the same thing and therefore can not be compared.
5.8 The brain has known operating frequencies. Input receptors fire twice a second. Motion is performed 10 times per
second. How does the brain accomplish the difference? A detailed explanation of this is contained in (Hempfling 1994)
and shall not be restated here. Suffice to say that the brain functions in the opposite of what would be considered to be
normal clock speeds.
5.9 From a single Biological Clock frequency operating frequencies are divided. The inputs are one frequency. The
subconscious comparisons based on long term memory are one frequency, the conscious comparisons and short term
memory are one frequency. The limbic and motion functions are one frequency. All derivatives of a single controlling
biological clock. (Hempfling, 1994.)
5.10 Intelligence occurs in the comparison process in both levels of it.
5.11 Consciousness occurs in the second level of comparison.
5.12 But how? By the differences in speeds of the levels. While the inputs are compared to long term memory in the first
or subconscious level of comparison those comparisons are made to long term memory that is running at a higher speed
than the inputs. Which results in mid term memory and subconscious intelligence that is greater than the inputs in count.
That means that what goes in must then be expanded. It is that expansion process that results in subconscious
intelligence. But it stops short at other awareness.
5.13 Other Awareness is the level of subconscious processing's result. It means the animal with this level of processing in
its brain will be aware of its predators. Aware of its rivals and therefore aware of its surroundings. But not aware of itself.
In Observational Illusion (Hempfling C 1995) man has observed this to be survival of the fittest when it is actually survival
of the most aware. The most efficient. It is not strength. It is intellect and capabilities to use that intellect.
5.14 In the human and some other lower forms of life in various stages of evolutionary development of intellectual
processing there resides a second comparator. The conscious level comparator compares the values of the mid term
memory (which is the result of the first level of comparison) to the values of the short term memory. The short term
memory once again functioning at a higher rate of speed than the mid term memory with the result of that second
conscious level comparison being sent to the short term memory setting up a feedback loop of mental computation. At
this point we can explain the quasi self awareness of lower than human creatures such as some forms of ape and
chimpanzee and others. As the short term memory is indeed that. A second or less of duration. Quasi-self awareness is
the result of the comparison of mid term to short term memory but it is lost to memory while outputted for motion in
lower quasi-conscious creatures as the feedback loop is so short in duration as to prohibit the concept of self to set up.
5.15 The concept of self in the human is not that far off and not that far advanced from the lower life forms. With a
human short term memory having a duration of a typical seven seconds a good deal of the brain is used. Short term
memory in the human is functioning at a much higher rate of computation than the lower life forms and that requires
short term memory even at just seven seconds to occupy a great deal of physical space. While mid term memory, being
results of the first comparison is far longer in duration (typical 18 hours,) (Hempfling 1995A) but far smaller in occupied
space.
5.16 Results of both levels of comparators is outputted to the collective area of the brain referred to as the Limbic
System. In the Limbic the speeds of computation and the multi paralleled computational pathways are blended since the
values of the comparisons get smaller with each level of computation. By the time they reach the Limbic the conscious
level is running quickly and with low values resulting in detailed control over motor functions as these values are blended
down and combined. While the subconscious level is running slower and with higher levels of values resulting in instant
and far less controlled motor functions as they are blended down and combined.
5.17 And finally if the reader has not already questioned it: Where does the memory value of the long term memory
come from? This is the major difference between humans and all other forms of life. In lower life forms the comparison of
the single comparator is sent to Limbic and long term. In higher forms of life it is sent from the second comparator but
with less frequency and less feedback the reduction performed from comparator to long term memory is less detailed
and results in memory that is off balance to input. In the human the speed of short term memory is so far advanced past
lower life forms and the values are so far reduced that the combination of them in the reduction to long term memory
results in a balanced value level to input. (Hempfling, 1995A) That means you can see an apple today and know an
apple tomorrow instantly where your cat will study it for a moment before becoming aware of the other. The apple. If
the apple is not important to the cat it simply will not bother with it. You on the other hand will bring up numerous
comparisons to the apple. As each visual and aural receptor's input pathway is split into three distinct parallel processing
pathways the depth of the comparisons is vast. And when you all of a sudden declare that you have had a brilliant flash
of intellect you can thank the apple. It compared to everything in recent memory that was the same shape, the same
color, the same reflection, the same flavor, the same smell, the same weight, the same temperature, the same angle, the
same everything. And each of those compared to other seemingly incomparable memories.
6. THE RAMIFICATIONS
6.1 If you live your life eating junk food your body will become junk. If you live your life inputting junk in the other inputs
to your body, your senses, your sense will become junk. If you are treated like junk when a very young child you will
treat others like junk. If all you see is junk as a young child all you will do will be junk. If you finally understand how the
brain does what it does you will be able to use it to remove the junk. Then you will not be any more intelligent than you
were before the knowledge. But you will be better equipped to evaluate others by knowing why they do what they do
the way they do it. And you will be better equipped to realize that as you see yourself is not as others see you. As you
do things for a reason it is not the reason others see you do it.
6.2 The real advantage to the understanding of intelligence and the process of the brain is not being able to make an
artificial variety. It is not in being able to bestow intelligence on a box of parts. It is in the utilization of the knowledge of
the process of thinking of all creatures in their various levels of intellect. It is treating others as you would have them treat
you. It is advancing past the level of evolution's gift of self awareness to once again realize the power of other
awareness. Something our friends on four feet have known all along. They just don't know they know it.
REFERENCES:
Baars, Bernard J.; Can Physics Provide a Theory of Consciousness? A Review of Shadows of the Mind by Roger Penrose.
PSYCHE: an interdisciplinary journal of research on consciousness
2(8), May 1995. Filename:psyche-95-2-08-shadows-6-baars ISSN: 1039-723X/93
Hempfling, Lee Kent; Why Digital Computers Will Never Be Intelligent, Internet Self Published (Copyright 1995)
http://members.aol.com/enticy1/ntc/select.htm
Hempfling, Lee Kent; (A) Understanding That Which Does The Understanding, Internet Self Published (Copyright 1995)
http://members.aol.com/enticy1/ntc/select.htm
Hempfling, Lee Kent; (B) On the Application of Spectroscopy to the Biological Vision Sensor. Internet Self Published.
(Copyright 1995) http://members.aol.com/enticy1/ntc/select.htm
Hempfling, Lee Kent; (C) Object Based Reasoning. Internet Self Published, (Copyright 1995)
http://members.aol.com/enticy1/ntc/select.htm
Hempfling, Lee Kent; The Biological Clock; Internet Self Published (Copyright 1994)
http://members.aol.com/enticy1/ntc/select.htm
McDermott, Drew; [STAR] Penrose is Wrong PSYCHE: an interdisciplinary journal of research on consciousness 2(17),
October, 1995. Filename: psyche-95-2-17-shadows-9-mcdermott
ISSN: 1039-723X/93
Merkle, Ralph C.; Large Scale Analysis of Neural Structures ; Xerox Palo Alto Research Center
CSL-89-10 November 1989, [P89-00173] Copyright 1989 Xerox Corporation.
Minsky, Marvin ; "Robotics," Anchor Press/Doubleday 1985
Smart, Bill; Artificial Intelligence, [email protected] 1994 Internet Self Published
