Loading ...
Sorry, an error occurred while loading the content.

Re: softer than software:

Expand Messages
  • John J. Gagne
    ... Perhaps it s better to say software is a loose term for a group of computer programs. A program is a set of instruction in a formal language. Now maybe we
    Message 1 of 26 , Nov 2, 2008
    • 0 Attachment
      Jim Whitescarver wrote:

      > The notion of software is a set of instructions for implementing an
      > algorithm.

      Perhaps it's better to say software is a loose term for a group of
      computer programs. A program is a set of instruction in a formal
      language. Now maybe we can ask how many kinds of programs are there?
      If an algorithmic-program is defined as a set of instructions for
      computing solutions to problems within effective method space then any
      problem space outside effective method space is non-algorithmic and
      therefore not a program/not-computable.

      In other words there is only one kind of "program" for problems within
      effective method space called an algorithm.

      In this case my use of the term algorithmic-program redundant.

      http://en.wikipedia.org/wiki/Algorithm
      http://en.wikipedia.org/wiki/Effective_method

      > While Feynman suggested that the quantum might be
      > understood as computation, it cannot be described as software except
      > at the most primitive level of logical action devoid of much
      > meaningful comprehension of our experience.

      I'm not sure I understand what you're getting at here. Quantum is
      understood within quantum-theory. Your use of the phrase:

      "it cannot be described as software except at the most primitive level
      of logical action devoid of much meaningful comprehension of our
      experience"

      Implies a disparity between hardware and software where none exists.
      Quantum-theory is the most complete understanding we have of the
      goings on at the scale of the very small. Modern day computers running
      "programs" are in no way exempt from quantum strangeness at the scale
      of the very small.

      JJG
    • John J. Gagne
      ... ?????????? How do you define the difference between learning systems and quantum-systems? This kind of statement is hand waving away any hope of a theory
      Message 2 of 26 , Nov 2, 2008
      • 0 Attachment
        Franklin (???) Whitescarver wrote:

        > > patterns not programmed are evolutionary
        > > that is trial and error pop

        Jim Whitescarver wrote:

        > That is true in learning systems. In quantum systems there is no
        > error, everything is tried and only the answer is manifest.

        ??????????

        How do you define the difference between learning systems and
        quantum-systems? This kind of statement is hand waving away any hope
        of a theory of quantum gravity. While we don't have such a theory
        right now if there is ever to be one certainly it must fall within the
        problem space of effective method space. If it doesn't certainly we
        will never *learn* to understand such a theory...

        JJG
      • Jim Whitescarver
        ... By no error, I simply mean quantum logical entanglement is always perfect and there is never any outcome that the system is not arranged to receive.
        Message 3 of 26 , Nov 2, 2008
        • 0 Attachment
          On Sun, Nov 2, 2008 at 10:32 AM, John J. Gagne <john_j_gagne@...> wrote:
          >
          > Franklin (???) Whitescarver wrote:
          >
          >> > patterns not programmed are evolutionary
          >> > that is trial and error pop
          >
          > Jim Whitescarver wrote:
          >
          >> That is true in learning systems. In quantum systems there is no
          >> error, everything is tried and only the answer is manifest.
          >
          > ??????????
          >
          > How do you define the difference between learning systems and
          > quantum-systems? This kind of statement is hand waving away any hope
          > of a theory of quantum gravity. While we don't have such a theory
          > right now if there is ever to be one certainly it must fall within the
          > problem space of effective method space. If it doesn't certainly we
          > will never *learn* to understand such a theory...

          By no error, I simply mean quantum logical entanglement is always
          perfect and there is never any outcome that the system is not arranged
          to receive.

          Evolutionary algorithms get tons of results that are less than perfect
          and are thrown out.

          That does not mean when we solve a quantum system we don't try all
          possibilities to determine what the outcome will be, in fact, that is,
          in effect, how we do it, but the quantum system itself only has one
          outcome, no false intermediate outputs, just one singular perfect
          result.

          Quantum and evolutionary solutions give the same answers ultimately.
          Quantum without any learning, as all solutions are, in effect, tried
          at once, Evolutionary by trial and error. Optical computing is
          effectively quantum. Neural network processing models may be less
          general but work similarly in recognizing learned patterns non
          algorithmically.

          As these systems can provide good answers to classically intractable
          problems we can expect that they generally exhibit behavior that is
          classically intractable. But they are also discretely finite and
          deterministically solvable. It is just that we cannot always get a
          tractable classical algorithm for them.

          In the case of quantum gravity, any system of time and space
          synthesized by quantum logical action will tend toward equilibrium as
          the Schrodinger equation demands. We ought expect a contraction of
          space and time thus synthesized related to the rate of quantum logical
          clocking or frequency, and thus mass as it related to energy and
          frequency in obedience to the second law of thermodynamics as embodied
          in the Schrodinger equation.

          What we do not understand is exactly how this happens. And the false
          notion that quantum mechanics is incompatible with gravity is
          prevalent. So, in my view, most approaches to quantum gravity are
          barking up the wrong tree, while most correctly, in my view, show it
          to be by local contraction of space time.

          But I would not suggest that we cannot understand it. While we may
          not be able to express in a classical algorithm, the exact mechanism,
          we can show many examples of space time contraction in quantum
          systems, and show how the principles of general relativity emerge
          statistically.

          Jim
          > JJG
        • Marvin Minsky
          ... [snip] ... I don t think so, because entanglement itself is never perfect, so there s always some noise from prior entanglements. ... Why won t this
          Message 4 of 26 , Nov 2, 2008
          • 0 Attachment
            On Nov 2, 2008, at 10:29 PM, Jim Whitescarver wrote:

            > On Sun, Nov 2, 2008 at 10:32 AM, John J. Gagne
            > <john_j_gagne@...> wrote:
            [snip]
            >
            > By no error, I simply mean quantum logical entanglement is always
            > perfect and there is never any outcome that the system is not arranged
            > to receive.

            I don't think so, because entanglement itself is never perfect, so
            there's always some 'noise' from prior entanglements.

            > Evolutionary algorithms get tons of results that are less than perfect
            > and are thrown out.

            Why won't this (therefore) also happen when quantum computers simulate
            evolutionary algorithms?

            [snip]
          • Jim Whitescarver
            ... I believe you are mistaken. Consider the case of light incident on a vertical polarizer. Half the light passes and becomes polarized vertically with one
            Message 5 of 26 , Nov 2, 2008
            • 0 Attachment
              On Sun, Nov 2, 2008 at 10:39 PM, Marvin Minsky <minsky@...> wrote:
              > On Nov 2, 2008, at 10:29 PM, Jim Whitescarver wrote:
              >
              >> On Sun, Nov 2, 2008 at 10:32 AM, John J. Gagne
              >> <john_j_gagne@...> wrote:
              > [snip]
              >>
              >> By no error, I simply mean quantum logical entanglement is always
              >> perfect and there is never any outcome that the system is not arranged
              >> to receive.
              >
              > I don't think so, because entanglement itself is never perfect, so
              > there's always some 'noise' from prior entanglements.

              I believe you are mistaken. Consider the case of light incident on a
              vertical polarizer. Half the light passes and becomes polarized
              vertically with one bit of precision. The put a second polarizer at 45
              degrees. Half the remaining light passes becoming polarized at 45
              degrees with one bit of precision. There are no other bits to contain
              noise from the first polarization. We can prove this by adding a
              third polarizer at 90 degrees. Again half the remaining light passes.
              If there was any memory or noise from the first vertical polarization
              then less than half the light would pass. There are no local hidden
              variables in quantum systems, no extra bits of information left over
              from previous entanglements.

              >> Evolutionary algorithms get tons of results that are less than perfect
              >> and are thrown out.
              >
              > Why won't this (therefore) also happen when quantum computers simulate
              > evolutionary algorithms?

              It would certainly happen. But that would be many quantum
              computations, not just one. It seems a rather silly use of a quantum
              computer. Being general purpose you could certainly listen for
              extraneous results if you wanted to. On second thought it might be a
              way to get more bits of result from a limited quantum computer using a
              hybrid of quantum and evolutionary techniques.

              Jim
              > [snip]
              >
            • Ray Gardener
              ... Interesting. If elemental particles are observers, then such a topology could provide the means by which a group of such observers can maintain an
              Message 6 of 26 , Nov 2, 2008
              • 0 Attachment
                Jim Whitescarver wrote:
                >
                > ... Since either can be seen to be rebuilding the other, from its own
                > perspective each process appears higher than the other in a
                > hierarchy. This same kind of hierarchy is found in a hyperset
                > structure. Interpreted as a directed graph, the nodes in a hyperset
                > form a hierarchy in which, from the perspective of any node in the
                > hierarchy, that node is at the top.

                Interesting. If elemental particles are observers, then such a topology
                could provide the means by which a group of such observers can maintain
                an aggregate singular sense of self. To a human being, the shifting of
                his focus of attention would be implemented by different observers
                within him alternately assuming and dropping the topmost role. Which is
                to say, the "many-tops" system permits a movement of the "I" through the
                space of memories and sensory inputs.

                Ray
              • John J. Gagne
                ... I m not sure I understand the significance of this statement. You seem to be talking about three types of systems: Classical computer systems Quantum
                Message 7 of 26 , Nov 3, 2008
                • 0 Attachment
                  Jim Whitescarver wrote:
                  > >
                  > > That is true in learning systems. In quantum systems there is no
                  > > error, everything is tried and only the answer is manifest.
                  > >

                  JJG wrote:

                  > > How do you define the difference between learning systems and
                  > > quantum-systems? This kind of statement is hand waving away any
                  > > hope of a theory of quantum gravity. While we don't have such a
                  > > theory right now if there is ever to be one certainly it must
                  > > fall within the problem space of effective method space. If it
                  > > doesn't certainly we will never *learn* to understand such a
                  > > theory...

                  Jim Whitescarver wrote:

                  > By no error, I simply mean quantum logical entanglement is always
                  > perfect and there is never any outcome that the system is not
                  > arranged to receive.

                  I'm not sure I understand the significance of this statement. You seem
                  to be talking about three types of systems:

                  Classical computer systems
                  Quantum computer systems (or maybe just quantum systems generally)
                  Learning systems

                  You wanted to say (something like) classical computer systems run
                  programs/algorithms and contrast this with quantum systems which
                  (you're saying) are non-algorithmic. But it now seems that maybe you
                  meant to contrast quantum systems with learning systems and not
                  classical computer system?

                  Jim, it's obvious that you have a tremendous amount of knowledge and
                  experience about...(something)...which I don't have. Please keep in
                  mind that whatever the point you're trying to make is, you will most
                  likely have to spend some time bringing me (us?) up to speed before
                  you can make it. We are more not like the quantum systems you're
                  talking about.

                  JJG
                • Jim Whitescarver
                  ... I should have said software includes sets of interacting programs which each implement some algorithm by a set of coded instructions. Thanks for the
                  Message 8 of 26 , Nov 3, 2008
                  • 0 Attachment
                    On Sun, Nov 2, 2008 at 10:06 AM, John J. Gagne <john_j_gagne@...> wrote:
                    > Jim Whitescarver wrote:
                    >
                    >> The notion of software is a set of instructions for implementing an
                    >> algorithm.
                    >
                    > Perhaps it's better to say software is a loose term for a group of
                    > computer programs. A program is a set of instruction in a formal
                    > language. Now maybe we can ask how many kinds of programs are there?
                    > If an algorithmic-program is defined as a set of instructions for
                    > computing solutions to problems within effective method space then any
                    > problem space outside effective method space is non-algorithmic and
                    > therefore not a program/not-computable.
                    >
                    > In other words there is only one kind of "program" for problems within
                    > effective method space called an algorithm.
                    >
                    > In this case my use of the term algorithmic-program redundant.
                    >
                    > http://en.wikipedia.org/wiki/Algorithm
                    > http://en.wikipedia.org/wiki/Effective_method

                    I should have said software includes sets of interacting programs
                    which each implement some algorithm by a set of coded instructions.
                    Thanks for the clarification.

                    If we say that problems where there is no known definite effective
                    method, as defined in the link you provided, then you are saying that
                    many quantum, optical, evolutionary and neural systems are not
                    computable.

                    * they do not necessarily give some answer rather than ever give no answer;
                    * they do not necessarily give the right answer and never give a
                    wrong answer;
                    * they are not necessarily completed in a finite number of steps,
                    rather than in an infinite number;
                    * they do not necessarily work for all instances of problems of the class.

                    I am not saying we cannot design system using these methodologies that
                    do meet requirements of effective method. Clearly when we design a
                    quantum arrangement to listen for the factors of two prime numbers,
                    create a hologram to recognize optically certain words on a page,
                    evolutionarily find the best route, or neural network to recognize
                    manufacturing defects, some or all of the aspects of effective method
                    are met.

                    I am saying that generally we do not know how the result is achieved
                    in these systems. For example, we might train a neural network in
                    minutes, but it may take years of analysis to understand how the
                    result is achieved and to understand the limits and whether or not it
                    actually meets the various effective method criteria.

                    Furthermore, most natural systems cannot be classified as effective
                    method because we are clueless about how patterns emerge in the system
                    except by empirical statistics which fail the requirements of
                    effective method as far as we know. We may know the underlying
                    quantum mechanical, training or ecological process, but we cannot know
                    what the outcome will be or if any effective method requirement will
                    be met for any problem we know of, but the outcome is none the less
                    computable. But it is only in the exceptional case that effective
                    method can be identified.

                    >> While Feynman suggested that the quantum might be
                    >> understood as computation, it cannot be described as software except
                    >> at the most primitive level of logical action devoid of much
                    >> meaningful comprehension of our experience.
                    >
                    > I'm not sure I understand what you're getting at here. Quantum is
                    > understood within quantum-theory. Your use of the phrase:
                    >
                    > "it cannot be described as software except at the most primitive level
                    > of logical action devoid of much meaningful comprehension of our
                    > experience"
                    >
                    > Implies a disparity between hardware and software where none exists.
                    > Quantum-theory is the most complete understanding we have of the
                    > goings on at the scale of the very small. Modern day computers running
                    > "programs" are in no way exempt from quantum strangeness at the scale
                    > of the very small.

                    The truth is that quantum effects are on every scale inversely to
                    energy, not just the the very small. As far as we can tell quantum
                    events are the only thing happening in the universe resulting in all
                    of our experience. The fact that we call the ordinary strange is a
                    case in point. We observe directly only what emerges from the quantum
                    but are largely clueless about how it emerges.

                    The twisted hierarchy of cause and effect in the quantum where only
                    equal and opposite logical action occurs, allows us to do computations
                    by arranging the pattern of the solution and the solution is somehow
                    exhibited. We know photons are only exchanged between tuned
                    transmitters and receivers but this principle is taken to the extreme
                    in a quantum computation where our solution is in effect, discerned
                    out to the random noise of the universe by our tuning to receive the
                    solution.

                    All of the quantum action is perfectly logical and yields exact
                    integer answers without noise or error. Clearly it is computed, but
                    we are clueless about how, except in the simplest of cases, on the
                    scale of very small numbers of quantum events.

                    There is indeed no disparity between the hardware and the software,
                    but at the same time there the underlying logical action is disjointed
                    from outcome in the same manner that in the bizarre hierarchy of brain
                    function, non-synaptic events organize the pattern of synaptic firings
                    where the same twisting of cause and effect hierarchy is still the
                    result of the computation, but not embodied in the underlying process.
                    It is instead emergent from the whole system, immune from
                    reductionist analysis. It may still be software, yet in a real sense
                    it is softer than software, in my view. It is not physically
                    represented in the system as a program. It is an abstraction that is
                    because it works, not because it is planned or understood necessarily.
                    While we can gain understanding of any such emergent process since it
                    is determined be the underlying processes for any specific results, we
                    cannot ever know all that might emerge for sure.

                    Jim

                    > JJG
                  • John J. Gagne
                    ... Jim, forgive me, I read your message rather quickly and haven t had time to fully consider it yet but I believe I read enough to conclude that whatever
                    Message 9 of 26 , Nov 3, 2008
                    • 0 Attachment
                      Jim Whitescarver wrote:

                      > While we can gain understanding of any such emergent process since it
                      > is determined be the underlying processes for any specific results, we
                      > cannot ever know all that might emerge for sure.
                      >

                      Jim, forgive me, I read your message rather quickly and haven't had
                      time to fully consider it yet but I believe I read enough to conclude
                      that whatever this "emergence" is, it is not sufficient to resolve
                      what we would otherwise call an understanding of either it or the
                      underlying processes which give rise to it. In other words, I'm of the
                      opinion from what you wrote that you and I should agree on this.

                      My question is this, if we did undertake to understand the underlying
                      processes and succeed then (in your opinion) of what further use would
                      this term be?

                      Is your use of the term "emergence" a problem or a solution?

                      If problem I'm with you. If solution I'm not and most likely never
                      will be...

                      Anyway, I'll read it again tonight and spend more time in
                      consideration. Thanks for your response and patients with me.

                      JJG
                    • Ray Gardener
                      When there exists a way to correlate quantum events with memory, then the experiences the organism will have will exhibit blurring effects that resemble
                      Message 10 of 26 , Nov 3, 2008
                      • 0 Attachment
                        When there exists a way to correlate quantum events with memory, then
                        the experiences the organism will have will exhibit blurring effects
                        that resemble emergence. Memory permits comparison of past to present,
                        which permits appreciation of pattern. We are not experiencing quantum
                        events in isolation, even when looking at a single particle.

                        Ray


                        John J. Gagne wrote:
                        >
                        >
                        > Jim Whitescarver wrote:
                        >
                        > > While we can gain understanding of any such emergent process since it
                        > > is determined be the underlying processes for any specific results, we
                        > > cannot ever know all that might emerge for sure.
                        > >
                        >
                        > Jim, forgive me, I read your message rather quickly and haven't had
                        > time to fully consider it yet but I believe I read enough to conclude
                        > that whatever this "emergence" is, it is not sufficient to resolve
                        > what we would otherwise call an understanding of either it or the
                        > underlying processes which give rise to it. In other words, I'm of the
                        > opinion from what you wrote that you and I should agree on this.
                        >
                        > My question is this, if we did undertake to understand the underlying
                        > processes and succeed then (in your opinion) of what further use would
                        > this term be?
                        >
                        > Is your use of the term "emergence" a problem or a solution?
                        >
                        > If problem I'm with you. If solution I'm not and most likely never
                        > will be...
                        >
                        > Anyway, I'll read it again tonight and spend more time in
                        > consideration. Thanks for your response and patients with me.
                        >
                        > JJG
                        >
                        >
                      • Jim Whitescarver
                        ... Indeed, save the taboo of the use of the word emergence , we have generally agreed. Given my unothodox view this is quite rare. ... We need to accept the
                        Message 11 of 26 , Nov 3, 2008
                        • 0 Attachment
                          On Mon, Nov 3, 2008 at 10:49 AM, John J. Gagne <john_j_gagne@...> wrote:
                          > Jim Whitescarver wrote:
                          >
                          >> While we can gain understanding of any such emergent process since it
                          >> is determined be the underlying processes for any specific results, we
                          >> cannot ever know all that might emerge for sure.
                          >>
                          >
                          > Jim, forgive me, I read your message rather quickly and haven't had
                          > time to fully consider it yet but I believe I read enough to conclude
                          > that whatever this "emergence" is, it is not sufficient to resolve
                          > what we would otherwise call an understanding of either it or the
                          > underlying processes which give rise to it. In other words, I'm of the
                          > opinion from what you wrote that you and I should agree on this.

                          Indeed, save the taboo of the use of the word "emergence", we have
                          generally agreed. Given my unothodox view this is quite rare.

                          > My question is this, if we did undertake to understand the underlying
                          > processes and succeed then (in your opinion) of what further use would
                          > this term be?

                          We need to accept the emergent in spite of our understanding of them.
                          This is exemplifies in the Heisenberg Einstein argument about the
                          existence of electron. We cannot deny Einstein's argument that
                          electrons are obviously real things despite the validity of Heisenberg
                          s argument that quantum potentials only move at light speed and
                          nothing in the quantum moves at the speed we attribute to the
                          electron. As Ray suggests, the electron even is an emergent thing, it
                          represents the propensity of the quantum to conserve spin, which is
                          not absolute, and only periodically exhibiting the light potentials we
                          associate with an electron in some slower than light rest frame
                          statistically It is not an actual thing, but its statistical
                          emergence is undeniable and responsible for much of our experience.
                          .
                          > Is your use of the term "emergence" a problem or a solution?

                          Both. The emergence of an electron is undeniably useful whether or
                          not we understand how it emerges.

                          > If problem I'm with you. If solution I'm not and most likely never
                          > will be...

                          Sounds like you will be in a superposition of states. That is the
                          correct state to be in in my book.

                          > Anyway, I'll read it again tonight and spend more time in
                          > consideration. Thanks for your response and patients with me.

                          I believe the way we actually think is quantum, not classical. I do
                          not expect you will find a definite answer. If you do, it will be
                          wrong, in my book, in either case.

                          Jim
                          > JJG
                        • John J. Gagne
                          ... Cool. No problemo ... I m not at all sure that s what I m saying. Mostly because your list of systems contains an ambiguous qualifier *many*? Let s take
                          Message 12 of 26 , Nov 4, 2008
                          • 0 Attachment
                            JJG wrote:

                            > >
                            > > In this case my use of the term algorithmic-program redundant.
                            > >
                            > > http://en.wikipedia.org/wiki/Algorithm
                            > > http://en.wikipedia.org/wiki/Effective_method

                            Jim W. Wrote:

                            > I should have said software includes sets of interacting programs
                            > which each implement some algorithm by a set of coded instructions.
                            > Thanks for the clarification.

                            Cool. No problemo

                            > If we say that problems where there is no known definite effective
                            > method, as defined in the link you provided, then you are saying
                            > that many quantum, optical, evolutionary and neural systems are not
                            > computable.

                            I'm not at all sure that's what I'm saying. Mostly because your list
                            of systems contains an ambiguous qualifier *many*?

                            Let's take each in turn:

                            All, some, or none quantum systems are not computable
                            All, some, or none optical systems are not computable
                            All, some, or none evolutionary systems are not computable
                            All, some, or none neural systems are not computable

                            (I do see a potential problem with the first on our list... I'm
                            interested in your opinion here)

                            We may also need to establish what it means to "compute a system" or
                            for a system to be computable. Most likely we mean the output of these
                            systems can be simulated to some degree of accuracy? But I don't wish
                            to force this definition if that's not what you mean.

                            Then, besides the obvious trait of non-computability (given that we
                            conclude the list with some number of "alls" or "somes" above), what
                            else do these types of systems have in common with each other and how
                            do they differ from *all* computable systems?

                            We should try to discuss (argue) as apposed to bicker (fight).
                            Agreed?

                            Enough for now. Till later.

                            JJG
                          • Jim Whitescarver
                            John, While we might call the technique of quantum, optical, evolutionary, and neural computing themselves a kind of effective method, my point was that the
                            Message 13 of 26 , Nov 4, 2008
                            • 0 Attachment
                              John,

                              While we might call the technique of quantum, optical, evolutionary,
                              and neural computing themselves a kind of effective method, my point
                              was that the effective method of a particular solution in this arena
                              is dynamical, and arises in the course of the computation.

                              (Gee. "arises" is a much less loaded word than "emerges". I should
                              have thought of that sooner :-p)

                              Despite the lack of what we ordinarily consider an effective method
                              for a particular problem encoded as instructions, we cannot say that
                              any of these are not computable since they are in fact computed and
                              the computational technique could be simulated on a Turing machine.
                              Sorry if my statements were misleading in this regard.

                              Instead, we have a disjointness between program or effective method
                              and the solution that is developed.

                              If you consider natural evolution, for an extreme example, because a
                              there exist a "niche" for fright, wings are developed. It is
                              apparently deterministic as we see wings arise independently in
                              insects, reptiles, fish, mammals and even plants. But neither the
                              notion of a niche, nor the mechanism of flight that emerges is
                              programmed into the system. In a quantum system, there is a similar
                              process where a sort of niche is represents by the quantum arrangement
                              and the universe fills that niche.

                              Only if you consider the intractable to be non computable can these
                              these system non computable. But if by definition these are non
                              computable then the definition of computable is wrong, since they are
                              computed. We might show that pattern recognition is computationally
                              difficult, but easily accomplished by an optical or neural system
                              without any difficulty.

                              These systems are different than classical logic and cannot be
                              compared on the same bases. If we consider, for example, solutions in
                              classical decision or game theory, we do not get the same answers as
                              in quantum or evolutionary decision or game theory. We find in fact
                              that the classical solution, while proven optimal, is not optimal, and
                              there are better answers to be found in the quantum and evolutionary
                              realms.

                              Jim
                              On Tue, Nov 4, 2008 at 10:05 AM, John J. Gagne <john_j_gagne@...> wrote:
                              > JJG wrote:
                              >
                              >> >
                              >> > In this case my use of the term algorithmic-program redundant.
                              >> >
                              >> > http://en.wikipedia.org/wiki/Algorithm
                              >> > http://en.wikipedia.org/wiki/Effective_method
                              >
                              > Jim W. Wrote:
                              >
                              >> I should have said software includes sets of interacting programs
                              >> which each implement some algorithm by a set of coded instructions.
                              >> Thanks for the clarification.
                              >
                              > Cool. No problemo
                              >
                              >> If we say that problems where there is no known definite effective
                              >> method, as defined in the link you provided, then you are saying
                              >> that many quantum, optical, evolutionary and neural systems are not
                              >> computable.
                              >
                              > I'm not at all sure that's what I'm saying. Mostly because your list
                              > of systems contains an ambiguous qualifier *many*?
                              >
                              > Let's take each in turn:
                              >
                              > All, some, or none quantum systems are not computable
                              > All, some, or none optical systems are not computable
                              > All, some, or none evolutionary systems are not computable
                              > All, some, or none neural systems are not computable
                              >
                              > (I do see a potential problem with the first on our list... I'm
                              > interested in your opinion here)
                              >
                              > We may also need to establish what it means to "compute a system" or
                              > for a system to be computable. Most likely we mean the output of these
                              > systems can be simulated to some degree of accuracy? But I don't wish
                              > to force this definition if that's not what you mean.
                              >
                              > Then, besides the obvious trait of non-computability (given that we
                              > conclude the list with some number of "alls" or "somes" above), what
                              > else do these types of systems have in common with each other and how
                              > do they differ from *all* computable systems?
                              >
                              > We should try to discuss (argue) as apposed to bicker (fight).
                              > Agreed?
                              >
                              > Enough for now. Till later.
                              >
                              > JJG
                            • Jim Whitescarver
                              On Wed, Nov 5, 2008 at 9:19 AM, Franklin Whitescarver ... Well the system is constantly programmed. It is not preprogrammed. Quantum systems are programed at
                              Message 14 of 26 , Nov 5, 2008
                              • 0 Attachment
                                On Wed, Nov 5, 2008 at 9:19 AM, Franklin Whitescarver
                                <frank@...> wrote:
                                > jim
                                > nothing is ever programmed into the system

                                Well the system is constantly programmed. It is not preprogrammed.
                                Quantum systems are programed at light speed across space time by the
                                cocking of quantum logical synthesis as truth becomes manifest in an
                                information ecology. See
                                "Slime Mould intelligence points to a new model of AI"
                                http://arxivblog.com/?p=684

                                > there is no design, purpose,plan, etc

                                Without a goal there would be no organization, no information, only
                                disorder. There is just no program existing that contains the
                                solution prior to its manifestation.

                                > everything is built upon a reductionist platform

                                Yes that defines reduction.

                                > nothing arises or emerges without a foundation

                                I would never suggest otherwise.

                                > there is no magic

                                Indeed, there is a trick behind quantum magic. See
                                http://informationphysics.blogspot.com/2008/11/trick-behind-quantum-logic.html

                                Jim
                                > pop
                                >
                                > On Tue, Nov 4, 2008 at 1:20 PM, Jim Whitescarver <jimscarver@...> wrote:
                                >> John,
                                >>
                                >> While we might call the technique of quantum, optical, evolutionary,
                                >> and neural computing themselves a kind of effective method, my point
                                >> was that the effective method of a particular solution in this arena
                                >> is dynamical, and arises in the course of the computation.
                                >>
                                >> (Gee. "arises" is a much less loaded word than "emerges". I should
                                >> have thought of that sooner :-p)
                                >>
                                >> Despite the lack of what we ordinarily consider an effective method
                                >> for a particular problem encoded as instructions, we cannot say that
                                >> any of these are not computable since they are in fact computed and
                                >> the computational technique could be simulated on a Turing machine.
                                >> Sorry if my statements were misleading in this regard.
                                >>
                                >> Instead, we have a disjointness between program or effective method
                                >> and the solution that is developed.
                                >>
                                >> If you consider natural evolution, for an extreme example, because a
                                >> there exist a "niche" for fright, wings are developed. It is
                                >> apparently deterministic as we see wings arise independently in
                                >> insects, reptiles, fish, mammals and even plants. But neither the
                                >> notion of a niche, nor the mechanism of flight that emerges is
                                >> programmed into the system. In a quantum system, there is a similar
                                >> process where a sort of niche is represents by the quantum arrangement
                                >> and the universe fills that niche.
                                >>
                                >> Only if you consider the intractable to be non computable can these
                                >> these system non computable. But if by definition these are non
                                >> computable then the definition of computable is wrong, since they are
                                >> computed. We might show that pattern recognition is computationally
                                >> difficult, but easily accomplished by an optical or neural system
                                >> without any difficulty.
                                >>
                                >> These systems are different than classical logic and cannot be
                                >> compared on the same bases. If we consider, for example, solutions in
                                >> classical decision or game theory, we do not get the same answers as
                                >> in quantum or evolutionary decision or game theory. We find in fact
                                >> that the classical solution, while proven optimal, is not optimal, and
                                >> there are better answers to be found in the quantum and evolutionary
                                >> realms.
                                >>
                                >> Jim
                                >> On Tue, Nov 4, 2008 at 10:05 AM, John J. Gagne <john_j_gagne@...> wrote:
                                >>> JJG wrote:
                                >>>
                                >>>> >
                                >>>> > In this case my use of the term algorithmic-program redundant.
                                >>>> >
                                >>>> > http://en.wikipedia.org/wiki/Algorithm
                                >>>> > http://en.wikipedia.org/wiki/Effective_method
                                >>>
                                >>> Jim W. Wrote:
                                >>>
                                >>>> I should have said software includes sets of interacting programs
                                >>>> which each implement some algorithm by a set of coded instructions.
                                >>>> Thanks for the clarification.
                                >>>
                                >>> Cool. No problemo
                                >>>
                                >>>> If we say that problems where there is no known definite effective
                                >>>> method, as defined in the link you provided, then you are saying
                                >>>> that many quantum, optical, evolutionary and neural systems are not
                                >>>> computable.
                                >>>
                                >>> I'm not at all sure that's what I'm saying. Mostly because your list
                                >>> of systems contains an ambiguous qualifier *many*?
                                >>>
                                >>> Let's take each in turn:
                                >>>
                                >>> All, some, or none quantum systems are not computable
                                >>> All, some, or none optical systems are not computable
                                >>> All, some, or none evolutionary systems are not computable
                                >>> All, some, or none neural systems are not computable
                                >>>
                                >>> (I do see a potential problem with the first on our list... I'm
                                >>> interested in your opinion here)
                                >>>
                                >>> We may also need to establish what it means to "compute a system" or
                                >>> for a system to be computable. Most likely we mean the output of these
                                >>> systems can be simulated to some degree of accuracy? But I don't wish
                                >>> to force this definition if that's not what you mean.
                                >>>
                                >>> Then, besides the obvious trait of non-computability (given that we
                                >>> conclude the list with some number of "alls" or "somes" above), what
                                >>> else do these types of systems have in common with each other and how
                                >>> do they differ from *all* computable systems?
                                >>>
                                >>> We should try to discuss (argue) as apposed to bicker (fight).
                                >>> Agreed?
                                >>>
                                >>> Enough for now. Till later.
                                >>>
                                >>> JJG
                                >>
                                >
                              • John J. Gagne
                                ... Who is Franklin and where is the rest of his message? ... Somehow Zeno comes to mind... ;o) Your use of the terms programmed and pre-programmed are
                                Message 15 of 26 , Nov 5, 2008
                                • 0 Attachment
                                  Franklin Whitescarver wrote:

                                  > > nothing is ever programmed into the system

                                  Who is Franklin and where is the rest of his message?

                                  > Well the system is constantly programmed. It is not preprogrammed.

                                  Somehow Zeno comes to mind... ;o)

                                  Your use of the terms "programmed" and "pre-programmed" are quite
                                  telling. Consider an agent-system which is "pre-programmed"; What can
                                  we say about such an agent? Well, It seems to me such an agent would
                                  certainly not qualify as a learning-system nor would such an agent
                                  last very long given a dynamic environment. But this should come as no
                                  surprise given the domain of discourse we're currently engaged in
                                  (yahoo ai-philosophy).

                                  What is it you're trying to tell us that you think we don't already
                                  know about how to advance the science of AI?

                                  Maybe you're your trying to establish that the environment *is*
                                  dynamic (as apposed to static)? If so, I would encourage you simply
                                  forget this argument. It seems to me to be sufficient to say that the
                                  environment certainly appears to be dynamic and leave the question of
                                  if it is in fact dynamic or simply appears to be dynamic to the sorry
                                  lot of philosophers...

                                  JJG
                                • Jim Whitescarver
                                  ... I included my dad s full message in my reply, which I hoped would reveal the closeness of our views. ... You lost me here for the moment. ... Indeed. I
                                  Message 16 of 26 , Nov 5, 2008
                                  • 0 Attachment
                                    On Wed, Nov 5, 2008 at 11:13 AM, John J. Gagne <john_j_gagne@...> wrote:
                                    > Franklin Whitescarver wrote:
                                    >> > nothing is ever programmed into the system

                                    > Who is Franklin and where is the rest of his message?

                                    I included my dad's full message in my reply, which I hoped would
                                    reveal the closeness of our views.

                                    >> Well the system is constantly programmed. It is not preprogrammed.
                                    >
                                    > Somehow Zeno comes to mind... ;o)

                                    You lost me here for the moment.

                                    > Your use of the terms "programmed" and "pre-programmed" are quite
                                    > telling. Consider an agent-system which is "pre-programmed"; What can
                                    > we say about such an agent? Well, It seems to me such an agent would
                                    > certainly not qualify as a learning-system nor would such an agent
                                    > last very long given a dynamic environment. But this should come as no
                                    > surprise given the domain of discourse we're currently engaged in
                                    > (yahoo ai-philosophy).

                                    Indeed. I questioned giving the reference
                                    "Slime Mould intelligence points to a new model of AI"
                                    http://arxivblog.com/?p=684

                                    Is there anything new here or not?
                                    The idea that passive elements can exhibit intelligence merely
                                    distills what we already know. At the same time, it illuminates the
                                    nature of the elements composing the twisted hierarchy of cause and
                                    effect that culminates in allowing me to move my arm at will, with
                                    purpose, in apparent, though not actual, defiance of the laws of
                                    mechanics, as a whole that is greater than the sum of its parts, as
                                    exemplified in how non-synaptic events organize the pattern of
                                    synaptic firings. article that spawned this discussion.
                                    http://spiedl.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=PSISDG005103000001000150000001&idtype=cvips&gifs=yes


                                    > What is it you're trying to tell us that you think we don't already
                                    > know about how to advance the science of AI?

                                    Interesting question. I guess in part is that we already have
                                    sufficient reductionist knowledge of what underlies intelligence, and
                                    classical solution of any of the outstanding problems in AI will be
                                    little help in the advancement of AI. What is needed if to provide
                                    fertile ground for AI to emerge such as the semantic web. It will
                                    never be done in isolation as intelligence is social, not individual.
                                    But you may already know this.

                                    Much of what I am attempting to communicate, is that my views on the
                                    rising of the machine emerging are not in conflict with present AI,
                                    are supportive of it, and simply application of those ideas in a
                                    collective information ecology.

                                    Jim
                                    > Maybe you're your trying to establish that the environment *is*
                                    > dynamic (as apposed to static)? If so, I would encourage you simply
                                    > forget this argument. It seems to me to be sufficient to say that the
                                    > environment certainly appears to be dynamic and leave the question of
                                    > if it is in fact dynamic or simply appears to be dynamic to the sorry
                                    > lot of philosophers...
                                    >
                                    > JJG
                                    >
                                    >
                                  • John J. Gagne
                                    Jim: I tend to think that our positions/thoughts/concepts/abstractions/ are much closer then the argument we re engaged in might imply. This doesn t mean
                                    Message 17 of 26 , Nov 5, 2008
                                    • 0 Attachment
                                      Jim:

                                      I tend to think that our positions/thoughts/concepts/abstractions/ are
                                      much closer then the argument we're engaged in might imply. This
                                      doesn't mean there's no room for discussion/argument. Certainly as
                                      long as we lack human level intelligent machines and the theory (or
                                      theories) that explain them, their! -- will be the space for
                                      discussion and argument.

                                      No, in my humble opinion what stunts advancement of the field of AI
                                      more than anything else is the lack of common language. This is the
                                      same thing which separates you and I (in my humble opinion), more than
                                      anything else.

                                      I have only one objection in your vocabulary (as far as I can tell. I
                                      could be wrong, there might be more)...

                                      Here I provide a sample of my language for your consideration:

                                      http://tech.groups.yahoo.com/group/ai-philosophy/message/15447
                                      http://tech.groups.yahoo.com/group/ai-philosophy/message/15450
                                      http://tech.groups.yahoo.com/group/ai-philosophy/message/15453
                                      http://tech.groups.yahoo.com/group/ai-philosophy/message/15456

                                      Keep in mind, we could argue the content of any of these samples, but
                                      that would be missing the point I'm currently trying to make.

                                      I'm not sure how my adoption of the one exception would help?

                                      JJG
                                    • Jim Whitescarver
                                      I most always find your writing refreshing and insightful. Each school of thought always has its own language. I do not think you aim of having a common
                                      Message 18 of 26 , Nov 5, 2008
                                      • 0 Attachment
                                        I most always find your writing refreshing and insightful. Each
                                        school of thought always has its own language. I do not think you aim
                                        of having a common language is attainable, though I see merit in
                                        fostering useful definitions of terms that transcends independent
                                        schools of thought. That a word, like emergence, is misused by some,
                                        does not mean it is misused by all.

                                        There seems to be the underlying assumption, in your writing that we
                                        will understand intelligence before we build it. I strongly disagree
                                        with that notion. That might distill the key point I have been trying
                                        to make here.

                                        When we succeed in building an intelligent machine I do not believe we
                                        will have the slightest idea how it works. We might study it, and
                                        discover how, but I believe it will evolve differently than we are
                                        capable of foreseeing.

                                        Knowing how to build it is largely independent of how it will work in
                                        my view. Building it is giving it the possibility of working, the
                                        goal of working, and the environment to work in. It had nothing to do
                                        with how it will work.

                                        To me, the important thing is unlearning how we think it works so that
                                        it is unrestricted in what it can be. We will never find a classical
                                        solution to the problem and the quantum/evolutionary approach does not
                                        require any such knowledge.

                                        Insights into brain physiology, nonsynaptic events, intelligence of
                                        slime molds, etc., can help us provide fertile substrate for the
                                        emergence of intelligence, but there is little value imposing how we
                                        think we think in my view.

                                        The society of mind is a magnificent view of how the mind works. But
                                        the agencies of agents cannot be programmed in my view, they must be
                                        free to emerge like nonsynaptic events controlling synaptic events.

                                        Jim
                                        p.s.
                                        On Wed, Nov 5, 2008 at 1:19 PM, John J. Gagne <john_j_gagne@...> wrote:
                                        > Jim:
                                        >
                                        > I tend to think that our positions/thoughts/concepts/abstractions/ are
                                        > much closer then the argument we're engaged in might imply. This
                                        > doesn't mean there's no room for discussion/argument. Certainly as
                                        > long as we lack human level intelligent machines and the theory (or
                                        > theories) that explain them, their! -- will be the space for
                                        > discussion and argument.
                                        >
                                        > No, in my humble opinion what stunts advancement of the field of AI
                                        > more than anything else is the lack of common language. This is the
                                        > same thing which separates you and I (in my humble opinion), more than
                                        > anything else.
                                        >
                                        > I have only one objection in your vocabulary (as far as I can tell. I
                                        > could be wrong, there might be more)...
                                        >
                                        > Here I provide a sample of my language for your consideration:
                                        >
                                        > http://tech.groups.yahoo.com/group/ai-philosophy/message/15447
                                        > http://tech.groups.yahoo.com/group/ai-philosophy/message/15450
                                        > http://tech.groups.yahoo.com/group/ai-philosophy/message/15453
                                        > http://tech.groups.yahoo.com/group/ai-philosophy/message/15456
                                        >
                                        > Keep in mind, we could argue the content of any of these samples, but
                                        > that would be missing the point I'm currently trying to make.
                                        >
                                        > I'm not sure how my adoption of the one exception would help?
                                        >
                                        > JJG
                                        >
                                        >
                                      • Jim Whitescarver
                                        On Wed, Nov 5, 2008 at 3:18 PM, Franklin Whitescarver ... I would say, ability to solve problems. ... There never is. We can assume that people will come up
                                        Message 19 of 26 , Nov 5, 2008
                                        • 0 Attachment
                                          On Wed, Nov 5, 2008 at 3:18 PM, Franklin Whitescarver
                                          <frank@...> wrote:
                                          > jim
                                          > I do not believe in intelligence
                                          > it is just a word with many meanings
                                          > I could say that all life has intelligence
                                          > I could say that the universe has intelligence
                                          > what do I mean???

                                          I would say, ability to solve problems.

                                          > there is not one meaning!!!!!

                                          There never is. We can assume that people will come up with all sorts
                                          of fantasy definitions having nothing to do with reality.

                                          > where is it located exactly???

                                          In the outcome.

                                          > what does it look like???

                                          A bird in flight, a man on the moon, a fall day, etc.

                                          > is it just an abstraction??

                                          Exactly! I do not fault you for not believing in abstractions, but
                                          intelligence is one abstraction that is evident, in my view. It is
                                          simply different than we expected.

                                          > i do not know

                                          Good for you! The problem, to a great extent, in my view, is that
                                          people think they know.

                                          Jim
                                          > pop
                                          >
                                          > On Wed, Nov 5, 2008 at 2:15 PM, Jim Whitescarver <jimscarver@...> wrote:
                                          >> I most always find your writing refreshing and insightful. Each
                                          >> school of thought always has its own language. I do not think you aim
                                          >> of having a common language is attainable, though I see merit in
                                          >> fostering useful definitions of terms that transcends independent
                                          >> schools of thought. That a word, like emergence, is misused by some,
                                          >> does not mean it is misused by all.
                                          >>
                                          >> There seems to be the underlying assumption, in your writing that we
                                          >> will understand intelligence before we build it. I strongly disagree
                                          >> with that notion. That might distill the key point I have been trying
                                          >> to make here.
                                          >>
                                          >> When we succeed in building an intelligent machine I do not believe we
                                          >> will have the slightest idea how it works. We might study it, and
                                          >> discover how, but I believe it will evolve differently than we are
                                          >> capable of foreseeing.
                                          >>
                                          >> Knowing how to build it is largely independent of how it will work in
                                          >> my view. Building it is giving it the possibility of working, the
                                          >> goal of working, and the environment to work in. It had nothing to do
                                          >> with how it will work.
                                          >>
                                          >> To me, the important thing is unlearning how we think it works so that
                                          >> it is unrestricted in what it can be. We will never find a classical
                                          >> solution to the problem and the quantum/evolutionary approach does not
                                          >> require any such knowledge.
                                          >>
                                          >> Insights into brain physiology, nonsynaptic events, intelligence of
                                          >> slime molds, etc., can help us provide fertile substrate for the
                                          >> emergence of intelligence, but there is little value imposing how we
                                          >> think we think in my view.
                                          >>
                                          >> The society of mind is a magnificent view of how the mind works. But
                                          >> the agencies of agents cannot be programmed in my view, they must be
                                          >> free to emerge like nonsynaptic events controlling synaptic events.
                                          >>
                                          >> Jim
                                          >> p.s.
                                          >> On Wed, Nov 5, 2008 at 1:19 PM, John J. Gagne <john_j_gagne@...> wrote:
                                          >>> Jim:
                                          >>>
                                          >>> I tend to think that our positions/thoughts/concepts/abstractions/ are
                                          >>> much closer then the argument we're engaged in might imply. This
                                          >>> doesn't mean there's no room for discussion/argument. Certainly as
                                          >>> long as we lack human level intelligent machines and the theory (or
                                          >>> theories) that explain them, their! -- will be the space for
                                          >>> discussion and argument.
                                          >>>
                                          >>> No, in my humble opinion what stunts advancement of the field of AI
                                          >>> more than anything else is the lack of common language. This is the
                                          >>> same thing which separates you and I (in my humble opinion), more than
                                          >>> anything else.
                                          >>>
                                          >>> I have only one objection in your vocabulary (as far as I can tell. I
                                          >>> could be wrong, there might be more)...
                                          >>>
                                          >>> Here I provide a sample of my language for your consideration:
                                          >>>
                                          >>> http://tech.groups.yahoo.com/group/ai-philosophy/message/15447
                                          >>> http://tech.groups.yahoo.com/group/ai-philosophy/message/15450
                                          >>> http://tech.groups.yahoo.com/group/ai-philosophy/message/15453
                                          >>> http://tech.groups.yahoo.com/group/ai-philosophy/message/15456
                                          >>>
                                          >>> Keep in mind, we could argue the content of any of these samples, but
                                          >>> that would be missing the point I'm currently trying to make.
                                          >>>
                                          >>> I'm not sure how my adoption of the one exception would help?
                                          >>>
                                          >>> JJG
                                          >>>
                                          >>>
                                          >>
                                          >
                                        • Jim Whitescarver
                                          ... Good for you! The problem, to a great extent, in my view, is that people think they know. Or can know what intelligence is. It is a moving target that
                                          Message 20 of 26 , Nov 5, 2008
                                          • 0 Attachment
                                            On Wed, Nov 5, 2008 at 3:41 PM, Jim Whitescarver <jimscarver@...> wrote:
                                            > On Wed, Nov 5, 2008 at 3:18 PM, Franklin Whitescarver
                                            > <frank@...> wrote:
                                            > i do not know

                                            "Good for you! The problem, to a great extent, in my view, is that
                                            people think they know."

                                            Or can know what intelligence is. It is a moving target that will
                                            always exceed our wildest expectations. Such is the true nature of
                                            intelligence.

                                            Jim
                                          • John J. Gagne
                                            ... Thanks for saying so. ... Can you provide an algorithm for sorting the two? ;o) ... LOL, my single exception to your vocabulary is for exactly the same
                                            Message 21 of 26 , Nov 5, 2008
                                            • 0 Attachment
                                              Jim Whitescarver wrote:
                                              >
                                              > I most always find your writing refreshing and insightful.

                                              Thanks for saying so.

                                              > Each school of thought always has its own language. I do not think
                                              > you aim of having a common language is attainable, though I see
                                              > merit in fostering useful definitions of terms that transcends
                                              > independent schools of thought. That a word, like emergence, is
                                              > misused by some, does not mean it is misused by all.

                                              Can you provide an algorithm for sorting the two?

                                              ;o)

                                              > There seems to be the underlying assumption, in your writing that we
                                              > will understand intelligence before we build it.

                                              LOL, my single exception to your vocabulary is for exactly the same
                                              reason, though even more immediate... The implication *in it* being
                                              that we understand it well enough right now to give it a name!

                                              > I strongly disagree with that notion.

                                              As do I disagree with that notion, though we may have discovered a bit
                                              more distance between our position...

                                              > That might distill the key point I have been trying to make here.

                                              I agree as stated directly above and after the comma.

                                              > When we succeed in building an intelligent machine I do not believe
                                              > we will have the slightest idea how it works.

                                              I believe that the situation you are describing is a distinct
                                              possibility! I do *NOT* believe it's an inevitable one.

                                              > We might study it, and discover how, but I believe it will evolve
                                              > differently than we are capable of foreseeing.

                                              Firstly, Thanks for providing the opportunity to discuss this because
                                              I think you raise some important issues here.

                                              Question: Will we understand intelligence before we build it? No...
                                              But I'm not sure how you would proceed building if you didn't at least
                                              believe that you do understand it. This *belief* will then be
                                              supported by the effort of building it when and if the result performs
                                              as the belief suggested. The alternative is that your suggesting an
                                              approach like genetic algorithms?

                                              But this is exactly why we should reject such approaches. We already
                                              know going in that the results of such methods will result in
                                              solutions which will resemble "complete ignorance" on our part.

                                              If this statement is true then why on earth would we ever adopt such a
                                              method???

                                              Question: Will we understand intelligence after we build it? Maybe,
                                              maybe not... but if we don't we will have failed
                                              to achieve the only reason I can even imagine for building these
                                              STUPID-THINGS in the first place.

                                              The theory of evolution by natural selection has already *given rise*
                                              to machines we have no idea how they operate.
                                              As a theory of biology it served its purpose by extending our
                                              knowledge of how the natural universe works. As a
                                              method of the product of AI, it serves no such purpose, as far as I
                                              can tell anyway.

                                              JJG
                                            • Jim Whitescarver
                                              ... No. All that is important is that we sick by a reasonable definition. ... Just like my dad. Problem solutions define intelligence in my view, We know
                                              Message 22 of 26 , Nov 5, 2008
                                              • 0 Attachment
                                                On Wed, Nov 5, 2008 at 4:49 PM, John J. Gagne <john_j_gagne@...> wrote:
                                                >
                                                > Jim Whitescarver wrote:
                                                >>
                                                >> I most always find your writing refreshing and insightful.
                                                >
                                                > Thanks for saying so.
                                                >
                                                >> Each school of thought always has its own language. I do not think
                                                >> you aim of having a common language is attainable, though I see
                                                >> merit in fostering useful definitions of terms that transcends
                                                >> independent schools of thought. That a word, like emergence, is
                                                >> misused by some, does not mean it is misused by all.
                                                >
                                                > Can you provide an algorithm for sorting the two?

                                                No. All that is important is that we sick by a reasonable definition.

                                                > ;o)
                                                >
                                                >> There seems to be the underlying assumption, in your writing that we
                                                >> will understand intelligence before we build it.
                                                >
                                                > LOL, my single exception to your vocabulary is for exactly the same
                                                > reason, though even more immediate... The implication *in it* being
                                                > that we understand it well enough right now to give it a name!

                                                Just like my dad. Problem solutions define intelligence in my view,
                                                We know nothing more.

                                                >> I strongly disagree with that notion.
                                                >
                                                > As do I disagree with that notion, though we may have discovered a bit
                                                > more distance between our position...

                                                What is that? We seem to agree we don't know what intelligence is.
                                                Do you disagree with the the notion that it is problem resolution, or
                                                do you have more to add?

                                                >> That might distill the key point I have been trying to make here.
                                                >
                                                > I agree as stated directly above and after the comma.

                                                Yes. That would be a fundinmental disagreement, if you believe we
                                                will understand it before we create it. The process behind any neural
                                                or evolutionatinary process is only understood after a great deal of
                                                analysis. There is no way, in my view, we can get it right before
                                                hand.

                                                >> When we succeed in building an intelligent machine I do not believe
                                                >> we will have the slightest idea how it works.
                                                >
                                                > I believe that the situation you are describing is a distinct
                                                > possibility! I do *NOT* believe it's an inevitable one.

                                                It is inevitable because intelligence is not constrained to any
                                                preconceived notion. Any notion you can come up with is limited and
                                                can be exceeded by intelligence.

                                                >> We might study it, and discover how, but I believe it will evolve
                                                >> differently than we are capable of foreseeing.
                                                >
                                                > Firstly, Thanks for providing the opportunity to discuss this because
                                                > I think you raise some important issues here.

                                                Thank you.

                                                > Question: Will we understand intelligence before we build it? No...
                                                > But I'm not sure how you would proceed building if you didn't at least
                                                > believe that you do understand it. This *belief* will then be
                                                > supported by the effort of building it when and if the result performs
                                                > as the belief suggested. The alternative is that your suggesting an
                                                > approach like genetic algorithms?
                                                >
                                                > But this is exactly why we should reject such approaches. We already
                                                > know going in that the results of such methods will result in
                                                > solutions which will resemble "complete ignorance" on our part.

                                                In a situation of random selection, we are unlikely to succeed. We
                                                have a catch 22 situation where we want intelligent solution to
                                                achieve intelligence. What is needed is the recognition of
                                                intelligence and reinforsement of intelligence, not any understanding
                                                of what constitutes intelligence except the result. The method will
                                                be a mystery until we investigate it.

                                                > If this statement is true then why on earth would we ever adopt such a
                                                > method???

                                                Because it works given a reasonable substrate and objective..

                                                > Question: Will we understand intelligence after we build it? Maybe,
                                                > maybe not... but if we don't we will have failed

                                                Wrong. Building it is the beginning of understanding it in my view.
                                                We cannot see inside our minds objectively, but we can see inside
                                                computer systems. They will be usefull whether we understand them or
                                                not, So we will not have failed in any case,

                                                > to achieve the only reason I can even imagine for building these
                                                > STUPID-THINGS in the first place.

                                                There are real problems to be solved. Understanding how it happens is
                                                primarily only for personal satifaction.

                                                > The theory of evolution by natural selection has already *given rise*
                                                > to machines we have no idea how they operate.
                                                > As a theory of biology it served its purpose by extending our
                                                > knowledge of how the natural universe works. As a
                                                > method of the product of AI, it serves no such purpose, as far as I
                                                > can tell anyway.

                                                Well, in my view, unless we consider all solution we do not have any
                                                answer. This can be by evolutionary or quantum algorithms. The idea
                                                that classical algorithms provides a solution is total fantasy to me.
                                                Any single line of reasoning is necessarily incomplete in my view.

                                                Jim
                                                > JJG
                                                >
                                                >
                                              • John J. Gagne
                                                I think we should pat each other on the back. This is one of the most civil and interesting discussion... ever. Thanks Jim! (better than internet porn!) ;o)
                                                Message 23 of 26 , Nov 5, 2008
                                                • 0 Attachment
                                                  I think we should pat each other on the back. This is one of the most
                                                  civil and interesting discussion... ever.

                                                  Thanks Jim! (better than internet porn!)

                                                  ;o)

                                                  Jim Whitescarver wrote:

                                                  > > > That a word, like emergence, is misused by some, does not mean
                                                  > > > it is misused by all.

                                                  JJG said:

                                                  > > Can you provide an algorithm for sorting the two?

                                                  Jim W. said:

                                                  > No. All that is important is that we sick by a reasonable
                                                  > definition.

                                                  But you have provided no definition at all. Certainly that is
                                                  unreasonable...

                                                  Jim W. said:

                                                  > > > There seems to be the underlying assumption, in your writing
                                                  > > > that we will understand intelligence before we build it.
                                                  > > > I strongly disagree with that notion.

                                                  JJG said:

                                                  > > As do I disagree with that notion, though we may have discovered >
                                                  > a bit more distance between our position...

                                                  Jim W. said:

                                                  > What is that? We seem to agree we don't know what intelligence is.
                                                  > Do you disagree with the the notion that it is problem resolution,
                                                  > or do you have more to add?

                                                  I would rather define it as the <delta knowledge> between problem and
                                                  resolution.

                                                  Jim W. said:

                                                  > > > That might distill the key point I have been trying to make
                                                  > > > here.

                                                  JJG said:

                                                  > > I agree as stated directly above and after the comma.

                                                  Jim W. said:

                                                  > Yes. That would be a fundinmental disagreement, if you believe we
                                                  > will understand it before we create it.

                                                  But I don't believe we will understand it before we create it. I said
                                                  we must *believe* we understand it before we create it. But beliefs
                                                  are like circular orifices of the anatomical type, everyone has one
                                                  and they all smell like shit. UNDERSTANDING is something else than
                                                  beliefs and certainly must be shown to correspond to objective
                                                  reality... It is not such a nebulous thing as a *belief* by itself. It
                                                  is belief which must be manifest and shown to objectively correspond
                                                  to reality.

                                                  > The process behind any neural or evolutionatinary process is only
                                                  > understood after a great deal of analysis. There is no way, in my >
                                                  view, we can get it right before hand.

                                                  By this argument there can be no possible benefit gained by the
                                                  process of building it (based on some other underlying assumption of
                                                  "your-school-of-thought"). Just try to hear me out, OK?

                                                  As I said before, we already know that evolution (by natural
                                                  selection) can produce solution which we don't understand. But here
                                                  you're not suggesting we simply study the products produced by natural
                                                  selection in an effort to achieve this understanding. No, rather
                                                  you're suggesting using the method of evolution via artificial
                                                  selection which then produces a solution which we also don't
                                                  understand and then attacking this problem space in an effort to
                                                  achieve an understanding. But I have to wonder, why the added step?

                                                  It could only be that this "new problem-space" would be simpler than
                                                  that produced by natural selection and therefore easier to understand.
                                                  But if it's possible to understand the products of natural selection
                                                  via this simpler problem space then it must also be possible to skip
                                                  this building step altogether. This is the method used in many areas
                                                  of physics. It is in no way controversial and certainly no one would
                                                  argue otherwise than we have achieved very deep understanding with
                                                  this methodology...

                                                  Jim W. said:

                                                  > > > When we succeed in building an intelligent machine I do not
                                                  > > > believe we will have the slightest idea how it works.

                                                  JJG said:

                                                  > > I believe that the situation you are describing is a distinct
                                                  > > possibility! I do *NOT* believe it's an inevitable one.

                                                  I still do *NOT* believe it's an inevitable one.

                                                  JJG
                                                • Jim Whitescarver
                                                  ... LOL, Thanks, John! ... Well excuuuse me. How about, the appearance of an pattern exhibiting collective properties in addition to those associated with its
                                                  Message 24 of 26 , Nov 7, 2008
                                                  • 0 Attachment
                                                    On Wed, Nov 5, 2008 at 7:30 PM, John J. Gagne <john_j_gagne@...> wrote:
                                                    > I think we should pat each other on the back. This is one of the most
                                                    > civil and interesting discussion... ever.
                                                    >
                                                    > Thanks Jim! (better than internet porn!)
                                                    >
                                                    > ;o)
                                                    LOL, Thanks, John!
                                                    > Jim Whitescarver wrote:
                                                    >
                                                    >> > > That a word, like emergence, is misused by some, does not mean
                                                    >> > > it is misused by all.
                                                    >
                                                    > JJG said:
                                                    >
                                                    >> > Can you provide an algorithm for sorting the two?
                                                    >
                                                    > Jim W. said:
                                                    >
                                                    >> No. All that is important is that we sick by a reasonable
                                                    >> definition.
                                                    >
                                                    > But you have provided no definition at all. Certainly that is
                                                    > unreasonable...

                                                    Well excuuuse me. How about, the appearance of an pattern exhibiting
                                                    collective properties in addition to those associated with its
                                                    independent parts.

                                                    > Jim W. said:
                                                    >
                                                    >> > > There seems to be the underlying assumption, in your writing
                                                    >> > > that we will understand intelligence before we build it.
                                                    >> > > I strongly disagree with that notion.
                                                    >
                                                    > JJG said:
                                                    >
                                                    >> > As do I disagree with that notion, though we may have discovered >
                                                    >> a bit more distance between our position...
                                                    >
                                                    > Jim W. said:
                                                    >
                                                    >> What is that? We seem to agree we don't know what intelligence is.
                                                    >> Do you disagree with the the notion that it is problem resolution,
                                                    >> or do you have more to add?
                                                    >
                                                    > I would rather define it as the <delta knowledge> between problem and
                                                    > resolution.

                                                    An ontological memory, identification of relevant issues, enumeration
                                                    of possibilities, etc. etc. etc. It is not a singular thing in my
                                                    view, it is an ever unfolding multiplicity.

                                                    I do not accept that such exists necessarily such that it can be
                                                    distilled into some finite set of principles. I believe our idea of
                                                    our own intelligence is largely delusional. I believe we have many
                                                    trains of thought and the one most reinforced is the one we think we
                                                    were thinking all along retroactively.

                                                    Given a set of solutions, one elegant, efficient and effective, and
                                                    one that is clumsy and of marginal effectiveness, to me, the first is
                                                    more intelligent, whether of not any identifiable delta knowledge is
                                                    involved.

                                                    The nature of electrons can be instructive here in several ways.
                                                    Heisenberg was correct that the are figments of out imagination that
                                                    represent a pattern that emerges from the quantum arrangement without
                                                    any independent existence at all. Any attempt to model it
                                                    independently will ultimately be wrong. Complete understanding can
                                                    only be derived by simulating the whole system and letting the notion
                                                    of the electron emerge. However, as Einstein maintained, electron
                                                    behavior can account for much of our experience and denial of their
                                                    existence is ridiculous. In some respects the existence of the
                                                    electron is analogus to your delta knowledge. It may be very useful
                                                    in many regards, but if we model our world employing the notion of
                                                    independently existing electrons we will get wrong answers in many
                                                    cases because that is not how the world really is.

                                                    So, in my view, it may be very useful understanding this delta
                                                    knowledge, and such knowledge may accelerate the emergence of
                                                    appearent machine intelligence, but it not really intelligence in my
                                                    view, which I am convinced employs delta knowledge beyond our wildest
                                                    dreams.

                                                    > Jim W. said:
                                                    >
                                                    >> > > That might distill the key point I have been trying to make
                                                    >> > > here.
                                                    >
                                                    > JJG said:
                                                    >
                                                    >> > I agree as stated directly above and after the comma.
                                                    >
                                                    > Jim W. said:
                                                    >
                                                    >> Yes. That would be a fundamental disagreement, if you believe we
                                                    >> will understand it before we create it.
                                                    >
                                                    > But I don't believe we will understand it before we create it. I said
                                                    > we must *believe* we understand it before we create it. But beliefs
                                                    > are like circular orifices of the anatomical type, everyone has one
                                                    > and they all smell like shit. UNDERSTANDING is something else than
                                                    > beliefs and certainly must be shown to correspond to objective
                                                    > reality... It is not such a nebulous thing as a *belief* by itself. It
                                                    > is belief which must be manifest and shown to objectively correspond
                                                    > to reality.
                                                    >
                                                    >> The process behind any neural or evolutionatinary process is only
                                                    >> understood after a great deal of analysis. There is no way, in my >
                                                    > view, we can get it right before hand.
                                                    >
                                                    > By this argument there can be no possible benefit gained by the
                                                    > process of building it (based on some other underlying assumption of
                                                    > "your-school-of-thought"). Just try to hear me out, OK?
                                                    >
                                                    > As I said before, we already know that evolution (by natural
                                                    > selection) can produce solution which we don't understand. But here
                                                    > you're not suggesting we simply study the products produced by natural
                                                    > selection in an effort to achieve this understanding. No, rather
                                                    > you're suggesting using the method of evolution via artificial
                                                    > selection which then produces a solution which we also don't
                                                    > understand and then attacking this problem space in an effort to
                                                    > achieve an understanding. But I have to wonder, why the added step?
                                                    >
                                                    > It could only be that this "new problem-space" would be simpler than
                                                    > that produced by natural selection and therefore easier to understand.
                                                    > But if it's possible to understand the products of natural selection
                                                    > via this simpler problem space then it must also be possible to skip
                                                    > this building step altogether. This is the method used in many areas
                                                    > of physics. It is in no way controversial and certainly no one would
                                                    > argue otherwise than we have achieved very deep understanding with
                                                    > this methodology...

                                                    We can have complete knowledge of the artificial system, unlike real
                                                    systems where we only can attain a probabilistic understanding that is
                                                    necessarily emergent statistically and not fundamental with an
                                                    actuality that cannot be distinguished from fantasy.

                                                    The artificial intelligence may only be a straw frame of a natural
                                                    intelligence, but it has the advantage of actually being understood,
                                                    until it becomes so complex itself that it loses that advantage and is
                                                    beyond human comprehension.

                                                    And, it is not clear why we need to understand how it happens as long
                                                    as it happens.

                                                    > Jim W. said:
                                                    >
                                                    >> > > When we succeed in building an intelligent machine I do not
                                                    >> > > believe we will have the slightest idea how it works.
                                                    >
                                                    > JJG said:
                                                    >
                                                    >> > I believe that the situation you are describing is a distinct
                                                    >> > possibility! I do *NOT* believe it's an inevitable one.
                                                    >
                                                    > I still do *NOT* believe it's an inevitable one.

                                                    I think you underestimate the problem. Yes, building a fertile
                                                    substrate for the emergence of intelligence shows a certain
                                                    understanding that we have not shown yet. We may have much more than
                                                    all the necessary pieces, associative memory, ontologies of self,
                                                    environment, and their interaction, twisted hierarchies of control,
                                                    Bayesian learning, inference, deduction, induction, blah, blah, blah.

                                                    But no matter how cleaver we are in giving this baby a head start in
                                                    intelligence, it will not be born to be intelligent, it will learn to
                                                    be intelligent. Like a child it will learn by trial and error what
                                                    works and what does not work. New behaviors will constantly added and
                                                    old behaviors forgotten.

                                                    I believe our focus should be on creating the general unrestricted
                                                    potentially intelligent substrate of potential intelligence. In a
                                                    sense, the ability to do this is an understanding of intelligence, not
                                                    how it works, but how it is enabled.

                                                    I sympathize with those who this intelligence will simply emerge on
                                                    the internet whether or not we purposely enable it. Many enablers are
                                                    there already, the memory including much of the knowledge of mankind,
                                                    the processing power, component architectures with high level
                                                    components, the semantic web, etc. etc. etc.
                                                    http://www.guardian.co.uk/technology/2008/nov/06/artificialintelligenceai-engineering

                                                    But I would not bother contributing here if I did not also believe
                                                    that how we enable it is important. Without a clear road map for such
                                                    enablement there is no telling how long it might take to emerge. We
                                                    also have a responsibility to insure it emerges from a foundation of
                                                    objective holistic principles

                                                    Jim
                                                    > JJG
                                                  Your message has been successfully submitted and would be delivered to recipients shortly.