If I understand Lee's comments correctly, I thank him for the apparent
approval of my intellect, and certainly agree with him on the need to
define intelligence before even talking about it (let alone undertaking
the building of an "intelligent" system). I certainly agree also on
the fundamental importance of proactive organization of input to a real
brain by that brain (it is not a tabula rasa, etc.).
I do have some reservations, however, about the reqirement that a truly
intelligent system must do it "as the brain does it", unless he has in
mind only the most basic and general requirement, i.e. proactive
structuring of input, without requiring that this be done exactly the
way the brain dopes it. Actually, I should quit saying "THE brain".
VERY early in my career (before many of you were born) I was
wonder-struck by (J. Zed?) Young's description of the octopus brain and
its radically different organization, making not only rat but even bird
brains seem by comparison similar to the human brain.
F. Frank LeFever, Ph.D.
New York Neuropsychology Group
In <199907052359.TAA30864 at triolog.net> lkent at MEDIAHUB.NET ("Lee Kent")
writes:
>>Dr. LeFever's comments are the basic intellect AI ignores.
>I base all such systems on their definition of intelligence. If they
>attempt to make intelligence without defining it first, then they are
>nothing more than actions, and the programs are nothing more than
>lists of organized rules for reactions. Worthless. Regardless of how
well an AI
>program does a task also performed by the brain, unless it does it as
>the brain does it, it is not intelligent let alone artificially
>intelligent. Intelligence is the level to which an entity is
>pro-active beyond the constraints and lessons learned that otherwise
>would result in nothing but reaction. There is no AI program that
>extends past reaction. CORE is pro-action. And true intelligence. Yet
>it is ignored by AI as it is based on real intelligence, something AI
>would not recognize if it was mandatory for survival.
>>Lee
>>On 5 Jul 99 at 19:13, Ron Blue wrote:
>>> > Seems to me I saw somebody's comment to the effect that AI systems
>> > do not HAVE to mimic natural systems and can stand on their own;
>> > but unless more novel/elegant/interesting as pure creations than
>> > these simplistic diagrams seem to imply, I see no point in
>> > pursuing such schemes even as a hobby or game...
>> >
>> > F. Frank LeFever, Ph.D.
>> > New York Neuropsychology Group
>> >>>>>
>> I would agree that AI systems do not have to mimic natural systems.
>> But a reduction to practice by using the natural components in
>> conscious biological systems in an AI environment has advantages.
>>>> For example, transistors can be used as self organizing maps
>> forming harmonic string memory in a linear trap of 45 years by using
>> a Paul trap frequency of 28 hertz. Classical confirmation of
>> information is required for harmonic self organizing memory maps
>> stored in oppositional reciprocal eigenfunctions. Because
>> transistors are quantum devices information can be self organized
>> instantaneously as long as the organization is confirmed
>> classically. Memory is local occurring as a perceptron or
>> oscillon and global stored in the linear trap of 45 years. Memory
>> is limited to 3.24 trillion qubits per transistor. Since the memory
>> is additive the global memory system is limited to the total number
>> of self organizing maps that can be formed in harmonic memory for
>> all the transistors in the system. Such a system is likely to be
>> superior to humans.
>>>> This models the correlational opponent processing or associational
>> reciprocal inhibition in biological systems. Biological systems
>> have oppositional wavelet information working at a Paul trap
>> frequency of 40 hertz. Biological systems use a linear time trap
>> of the life of the individual.
>>>> Ron Blue
>>http://www.neutronicstechcorp.com>>>>>>>>>>>Lee Kent - UCHubNet Internet Services
>lkent at uchub.net; webmaster at uchub.net>webmaster at triolog.net; lee at leekent.com>http://www.uchub.net>Get 7.9 All The Time.. sign up for UCHub
>Long Distance.. http://www.uchub.net/home/signup.html
