Can artificial intelligences suffer from mental illness?

Discussion in 'Intelligence & Machines' started by Plazma Inferno!, Aug 2, 2016.

  1. Write4U Valued Senior Member

    Messages:
    18,777
    You're right if that were the case, but all sensory organs have specific neural connections, which are directly routed to the appropriate section of the brain, ie. colors of light are received by the eye and must pass through the optical neural system, which translates and sends the information directly to the "visual department".

    The term "neural network" does not exclude specialized neurons which are directly connected to our 4 fundamental receptors, each adapted as efficient transmitters of specific signals. They are all part of the system.

    Look at the variety of cables used in a computer, each occupying a port which only translates specific information in the ever more efficient (human designed) configuration.
    https://en.wikipedia.org/wiki/Sensory_nervous_system

    I believe that's the idea, except it is not a deliberate action. The brain does all the work as efficient as it is able.
    What makes you think the source of sound cannot be approximated. We possess stereo vision, stereo hearing, stereo smell and by the use of two arms to touch, we achieve 3D touch.

    All of them can work individually or in tandem, depending on the various sources of origin, tocome to a best guess of what is being consciously experienced.

    As always I keep coming back to the "mirror neuron system" which is the processor of input from the external world and offers cognition by association and recognition from memory of previous associations.
    The sections where memories of experiences are stored and occasionally accessed.

    This would account for gradual memory loss of memories a long time ago. They are being replaced by newer memories.
     
    Last edited: Feb 6, 2018
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Michael 345 New year. PRESENT is 72 years oldl Valued Senior Member

    Messages:
    12,856
    Did you miss the part where I am describing VISION inputs arriving at the AUDIO region - Mr Audio tech does NOT redirect to correct region "Hey Mr Vision these are yours" he just merrily treats the VISION impulses he receives as AUDIO

    If that sort of mix occurred between the hands and feet we could spend part of our life walking on hands trying to type with our feet

    Don't tell me I don't know what I am doing. I know that. But I have no way I can tell my brain "Hey seperation of impulses. If I see something I want to SEE it, not HEAR it. Get your/my act together"

    Please Register or Log in to view the hidden image!

     
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. iceaura Valued Senior Member

    Messages:
    30,994
    The object - any object - comprises a set of interactions between - - - - - - between what, become the questions. As far as humans can describe, apparently, it's objects within objects all the way down - and up. Don't forget up: we're just as lacking in good description of the very complex large as we are of the interrelating very small.
     
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. Michael 345 New year. PRESENT is 72 years oldl Valued Senior Member

    Messages:
    12,856
    Granted

    The point was about interactions being computable

    We MIGHT be able to compute actions between detectable things if the number of things is kept to a manageable size, no matter the size of the objects, and the computation does not extend to far into the future

    To far into the future and you start getting into uncertain chaos and butterfly effects

    Please Register or Log in to view the hidden image!

     
  8. Write4U Valued Senior Member

    Messages:
    18,777
    Visual inputs have no effect on the auditory system , it is like asking a microphone to receive and transmit photons to the amplifier. But a microphone only can receive Audio waves, which are air pressure waves.

    OTOH, a camera is unable to process Audio pressure waves, but instead are sensitive to EM waves, and a very limited range at that .

    IMO, there are different kinds of receptors which are specific to the medium they have evolved to process but are not sensitive to other media which may also be present within the area of perception.

    IOW, when a bomb explodes, our eyes will see the bright flash, but our ears will only hear the loudness of the explosion, and if it is a big bomb, our bodies may detect a shock wave or a trembling of the ground.

    From these three separate sentient experiences and comparing them to prior experiences, our brains are able to process and integrate the separately processed information, creating an inner best guess that what we are experiencing are the result of an exploding bomb.

    Consider this; a blind person may hear the explosion and physically feel the shockwave, but will not experience the accompanying flash of light. OTOH, a deaf person may observe the flash and feel the shockwave, but will be unable to hear the sound.
     
  9. Michael 345 New year. PRESENT is 72 years oldl Valued Senior Member

    Messages:
    12,856
    Synesthesia

    Would seem to dispell that view

    *****
    For Duke Ellington, a D note looked like dark blue burlap while a G was light blue satin. When Pharrell Williams listened to Earth, Wind & Fire as a kid, he saw burgundy or baby blue. For Kanye West, pianos are blue, snares are white, and basslines are dark brown and purple. Orange is a big one for Frank Ocean.

    All of these artists—along with Stevie Wonder, Billy Joel, Mary J. Blige, Blood Orange's Dev Hynes, and more—have synesthesia, a condition in which a person's senses are joined.


    https://pitchfork.com/thepitch/229-...-and-why-does-every-musician-seem-to-have-it/

    *****

    Please Register or Log in to view the hidden image!

     
  10. Michael 345 New year. PRESENT is 72 years oldl Valued Senior Member

    Messages:
    12,856
    More on synesthesia

    Stephen

    . . . saw . . . there's Rimsky-Korsakov .. . he saw it as a . . . as a glorious sunny yellow. Er . . . Liszt would say to orchestras, and completely baffle them, erm, "No, no, no, please, gentlemen: bluer, bluer." Julian Asher, who is a neuroscientist, who also has synaesthesia, he tried to explain it, 'cause he had it as a child, and he used to get taken to concerts by his parents, and he always used to assume that the lights went down before the concert so that you could see the colours better as they came off the orchestra. He just always assumed that--'cause he assumed, as you would, that everybody had the same experience--that when they heard music, they saw colours, right in front of them, for real. And . . . and, er, Rimsky-Korsakov wrote down what he saw, and so, going up from F, we have E major--[brings right hand down as E major chord plays]--er, and that for him was bright blue. And F major--[brings right hand down as F major chord plays]--was . . .

    Alan
    Red.

    Stephen
    Ah. Bright green for him. But this is interesting, E-flat major--[brings right hand down as E-flat major chord plays]--er . . .

    Linda
    Magnolia.

    Stephen
    Miserable grey for him. Well, isn't that interesting? 'Cause it's very, very common indeed, E-flat major, for singers. It's . . . it's their most common chord

    QI TV program transcrips

    https://sites.google.com/site/qitranscripts/transcripts/2x11

    Please Register or Log in to view the hidden image!

     
  11. Write4U Valued Senior Member

    Messages:
    18,777
    As an ex musician I can associate certain notes or chords with colors or hues and an additional range of internal emotional associations created by my brain. I love the "Blues"

    But that is not a result of hearing the color "blue" or seeing the wave frequency of the note "a". It is an internal neural association within the brain, which in music theory is developed along with the concept of "shading" sounds.

    Some trumpet players play the instrument with a "velvety" tone, others play with a "sharp edge". Can you emotionally experience that?

    Phil Spector created a "wall of sound". Can you visualize that?

    Minor chords sound "sad", major chords sound "happy". Can you experience those emotional associations?

    In music, effective arrangements (or improvisations) are those that build emotional "tension" followed by emotional "release".

    IMO, a clear case of associative cognition

    When I said each neural transmission system has a specific and separated purpose, does not mean that brain itself is not able to assemble abstract similarities or metaphorical experiences from otherwise unrelated phenomena processed in various dedicated parts of the brain. By necessity all those functions are connected and work in tandem, in order to make a best guess or approximation of the received input.
     
    Last edited: Feb 7, 2018
  12. someguy1 Registered Senior Member

    Messages:
    727
    I'm falling behind in my replies but I'm slogging forward. Will get to everything eventually. I'm finding this thread very clarifying for my own thoughts.

    Yes of course. Substrate independence. The particular technique of implementing a computation doesn't matter. You can do it on a supercomputer, with pencil and paper, with a neural net, with a quantum computer. Any computation is reducible to a TM and TMs are independent of their physical implementation. "Methods and means" do not matter.

    Perhaps this is a good time to note the difference between computability theory and complexity theory. In computability theory, everything's a TM and that's that. In complexity theory we do care about how fast algorithms run. It's entirely possible that the secret to mind and the universe does lie in complexity theory rather than mere computability. Some people think that. That's the quantum computer argument. Quantum computers are computationally equivalent to TMs, but may be stronger in terms of complexity theory. For example there's a quantum algorithm to factor integers in polynomial time. There's no known algorithm to do the same on a conventional computer. I believe it's an open question as to whether QCs are more powerful than TMs in the sense of complexity theory.

    LOL. Yes I did choose a bad example. Smart quotes are equivalent to the halting problem! (jk)

    That's an interesting comment and I'll try to unpack it a bit.

    First, you stated the unsolvability of the halting problem inaccurately. You said that the halting problem is unsolvable "in principle." That's far too strong a statement. Turing proved that no computation can solve the halting problem.

    But nobody knows whether the halting problem may in fact be solvable by humans! If we could solve the halting problem, that would show that we're not computations. It's an open problem. I found a fascinating Stackexchange thread on the question.
    https://cs.stackexchange.com/questi...decide-the-halting-problem-on-turing-machines

    That's a minor issue. More importantly, you're muddying the point a bit IMO by bringing in non-halting computations.

    Consider the following problem. I give you the exact position and velocity of every particle in the universe at time t-zero. I ask you to determine the exact position and velocity of each particle at time t-one, which is millions of years in the future.

    We are already fairly sure (but not absolutely certain as far as I can tell) that if the physics is Newtonian, this can not be done. But if we use quantum physics, the question is open, and very important.

    But please note that the rules of this game require that the computation must consist of a finite number of instructions; and that it must halt on the correct answer within a finite number of steps. Those two conditions are essential to what we call computability.

    I hope you can see that the any of the following three states of affairs would be wholly unacceptable as a solution to the problem I posed:

    * Halting after a finite number of steps, giving no answer at all (an error condition, or a meaningless response); or

    * Halting after a finite number of steps, giving the wrong answer; or

    * Not halting.

    If the algorithm never halts, it never gives any answer at all.

    In short, in order for us to claim some problem is computable, it must be able to be computed by a TM whose program consists of finitely many instructions, each of which has finite length; and it must terminate (giving the correct answer) within a finite number of steps.

    That's not some guy on the Internet (me) saying that. It's what the people who study computations say. It's true that there's an extensive literature on infinitary computations, beginning with Turing's own doctoral thesis. But these models do not pertain to the physical world according to currently accepted theories of physics.

    I'm afraid I didn't understand that remark in the context of the present discussion.

    Every particle is exactly where it is at any given moment of time.

    Of course we are now bumping up against a deep question: Whether quantum uncertainty is an ontological or an epistemological phenomenon. That is, is quantum uncertainty a feature of nature itself, where the laws of physics "play dice" to determine whether Schrodinger's cat is dead or alive? Or is it merely that we lack the ability to know which quantum choices will be made? It's an open problem for philosophers and physicists.
     
    Last edited: Feb 7, 2018
  13. someguy1 Registered Senior Member

    Messages:
    727
    Thanks for the pointer. I'm reading up on slime mold intelligence. Very interesting.
     
  14. Write4U Valued Senior Member

    Messages:
    18,777
    A perfect example of associative thinking. Each person experienced a different color from the same auditory information, but by different internal emotional association.

    If we were to show the actual color blue to each of the persons, they would all agree that they were looking at the color blue.

    Now that in itself might trigger an association with other sensory abilities. Some people can associate the color blue with experiencing the taste of salt. Perhaps from the combinatory experience of living near the ocean and seeing the blue water and tasting the salty air.
    This sensory experience of this environmental condition is recorded in memory and later may possibly be recalled when seeing the color blue.

    If anything, IMO, this argues against randomness in the thought processes of the brain, but seems to confirm the process of trying to associate one sensory experience with another prior experience.
    The process is inexact, to be sure. But that's why Anil Seth calls it "best guess".
     
  15. someguy1 Registered Senior Member

    Messages:
    727
    Don't know. Some people these days claim that the mind and/or the universe are computations. I'm simply applying the basics of the known theory of computations to analyze that claim. Personally I'm not convinced there even are natural laws, or that if there are, that we could ever know them.
     
  16. someguy1 Registered Senior Member

    Messages:
    727
    If you believe that the universe operates according to laws and that these laws are not computational, then you join me in disagreeing with those who claim the world is a computer. You don't need to try to convince me of that! On the other hand if there are rules, then why aren't the rules computational? What are these rules doing that goes beyond our present notion of computation? Church-Turing again.
     
  17. Michael 345 New year. PRESENT is 72 years oldl Valued Senior Member

    Messages:
    12,856
    I can appreciate the emotional association of colours with sounds and I understand that

    BUT

    the situation below, and from a neuroscientist is not just an association

    These are associations (from the same episode's transcript)

    https://sites.google.com/site/qitranscripts/transcripts/2x11

    Stephen
    Exactly. Good. Now, listen to this piece of music by The Mamas and The Papas.

    [The song "Monday Monday" by The Mamas and the Papas plays on speakers.]

    Stephen
    Lovely. That, as I say, is the sound of "Monday Monday" by The Mamas and the Papas. But what colour would you say Monday is?

    Alan
    [presses buzzer, which plays second bar of 'Twinkle Twinkle Little Star', quickly, on a glockenspiel]

    Stephen
    Yes.

    Alan
    Blue.

    Stephen
    Blue. Because of . . . why? Because . . . ?

    Alan
    It just makes me think of blue.

    Stephen
    Well, that's right. Most people, if they think of days of the week, assign a colour to them. Not "assign" one, . . .

    Alan
    [quickly] Wednesday is kind-of green, Thursday is brown, Friday is black.

    Stephen
    Ah, you see, Friday is dark . . . dark blue to me, and Thursday is sort of orange, a red-y, beet red . . .

    Alan
    Tuesday is, maybe, yellow.

    Stephen
    --yellow. I have a yellow Tuesday.

    Alan
    Saturday's red.

    Stephen
    Yeah.

    Alan
    Sunday's . . . sort-of blue-y purple.

    Stephen
    Yes. Monday's white to me, for some reason. But there you are.

    Sean
    Have you done a spider experiment at NASA, you two?

    Stephen
    Do you not have any sense of colour when you think of . . .

    Sean
    No, I always think of . . . Monday, I think of a period of time which has to be endured until Tuesday comes along.

    Stephen
    Yes, don't . . . don't expect the Poet Laureate-ship to be handed to you on a plate! It's a little bit literal.

    Alan
    February is yellow.

    Linda
    If Monday makes you think of blue, does blue make you think of Monday?

    Alan
    It makes me think of mould.

    Stephen
    Yeah. It's not "thought", either. I mean, that is to say, it's not thought in the sense of rational analysis. It's as if you see the colour in your head. Well, we'll move onto this because, er, I'll play you a chord of music like so. [brings right hand down into the air, as a piano chord simultaneously plays] Which some of you with perfect pitch might know was D major . . . D, F-sharp, A triad.

    Alan
    Lime green, I had there.

    Stephen
    You had lime green? Anybody else have a colour?

    Linda
    [pause] No!

    Stephen
    No?

    Linda
    You know what I heard? I heard a sound.

    Sean
    I just thought . . .

    Stephen
    Yes. I know. I know. We do . . . most people . . .

    Linda
    [to audience] Anyone hear a sound? Anyone? Or just me? Please, God, not again!

    [A few members of the audience applaud in response.]

    Sean
    I just thought of, also, our sound man going, "Yes!"


    Please Register or Log in to view the hidden image!

     
  18. Michael 345 New year. PRESENT is 72 years oldl Valued Senior Member

    Messages:
    12,856
    Best response I can think of just off the top of my head

    if you had a Universe without LIFE yes you stuff obeying the Laws of Physics in which the actions / reactions would be computable

    LIFE in the Universe with brains (in the case of Earth 7+ billion human one's) which I contend are NOT computable and which affect the Universe in unpredictable ways cause the predictability of the Universe to become unpredictable

    Any programming buffs out there who can write a large PROGRAM with a 100% predictability of the output? Call it PHYSICS
    Can you then have a separate PROGRAM which throws random size inputs in (NOT changing the VALUES of the ridged ones of PHYSICS) at RANDOM intervals? Call it BRAIN

    My guess is that you will NOT get out predictable PHYSICS results

    Thanks for the reply as I think you helped me clarify what I had been pondering

    Weither it makes sense to you??? Well?

    Please Register or Log in to view the hidden image!

     
  19. Q-reeus Banned Valued Senior Member

    Messages:
    4,695
    A strong indicator of useless prattle is liberal use of CAPITALIZED words as poor substitute for actual well reasoned argument.
    Like trying to bolster via bluster a 'theory' that, notwithstanding QM, a formally perfectly predictable universe goes haywire, when 'Life -> Brains -> Laws of Physics Defying, Intrinsic Randomness'. Such a 'theory' being utterly devoid of objective, testable evidence.
    Still it must be admitted random input is a key feature here at almost-anything-goes SF.
     
  20. iceaura Valued Senior Member

    Messages:
    30,994
    Every particle is also a wave, presenting much the same problems in defining its location as the ripple in the coffee in the first place. And whatever definition or convention one eventually adopts for the "exact" location raises the question of whether the universe's computations are employing the same one.
    I didn't state it at all. I stipulated to its irrelevance.
    The universe does not have to "solve" the halting problem to run as we see it run, whether or not it is - in principle - "computable".
    The only reason it has to halt in a finite number of steps is that human engineering limitations require each step to take a certain amount of time. That would not apply to a universe that is producing time itself via its computations, and taking its steps on a rational number line. (Note: countability preserves computability, in principle - the rationals can be put into 1-1 correspondence with the integers).
    The universe may very well be capable of taking its steps in infinitesimal "time", and with infinitesimal spacing. At least, there is no obvious reason to assume otherwise, eh?
    You are dealing with a version of Zeno's Paradox, here. The arrow does move.
    Exact positions, velocities, and times, probably do not exist as properties of "particles". More likely, like temperatures and densities, they are human abstractions and heuristics whose usefulness emerges at appropriate scales of inquiry.
     
    Last edited: Feb 7, 2018
  21. someguy1 Registered Senior Member

    Messages:
    727
    In other words, life is the universe's solution to the halting problem. It's life that allows the universe to go beyond the Turing machine. I think that's very reasonable. Whether it's true, well who knows.
     
  22. Michael 345 New year. PRESENT is 72 years oldl Valued Senior Member

    Messages:
    12,856
    Whether true or not I am not smart enough to know or work out

    I did think I could test throwing random numbers into a fixed running program (Physics) without changing the fixed values of the running program but making it swerve away from its previously set destination outcome.

    Throw enough random numbers in would produce enough swerves to essentially render the fixed program indistinguishable from the random program

    Please Register or Log in to view the hidden image!

     
  23. someguy1 Registered Senior Member

    Messages:
    727
    I certainly agree with you that quantum physics says that particles aren't exactly where they are ... rather, each particle is a probability wave of where in the universe it might be.

    I did mention earlier that this is the philosophical problem of whether quantum randomness was an ontological or an epistemological problem. Is quantum randomness a feature of the universe? Or is it only a limitation of our knowledge?

    I think if nothing else, it's clear that the current state of physics doesn't allow anyone to say with certainty whether the mind is a computation or not. I don't see how anyone could claim certainty of anything. Yet so many people do.

    I don't follow this. Nobody said the universe solves the halting problem. I never said that. The entire point is that IF the universe is a computation, THEN it cannot solve the halting problem.

    But frankly I never brought up the halting problem. You did. If I misunderstood your intention, my apologies. That said, I don't understand how the remark you made bears on anything I said. Perhaps we're talking past each other. I think we're mostly in agreement.


    Not true. The reason is that our knowledge of fundamental physics does not allow us to physically instantiate infinitary models of computation.

    When I say that a computation is required to halt in order for its output to be called a computable object (real number, state of the universe, whatever), that is not my opinion. It's the official technical definition. It drives me a little nuts when you make up your own science. The Wiki links are all out there. It's like if I said gravity is an attractive force and you say no it's not, it's repulsive. You don't get to say that. Because it's already an established technical term in physics. Likewise, what it means for an object to be computable. It means that this object is the output of a Turing machine program that halts after finitely many steps.

    This isn't a matter of opinion.

    We can certainly talk about infinitary models of computation. They are very interesting and there are a number of different approaches. But none of them are implementable in the physical world according to our contemporary physics.

    If you want to agree with me that we need new physics, then we're in agreement. But if you want to argue with me that current physics allows infinitary computation, that's just wrong.

    I hope at least I've made my position as clear as I possibly can. If we disagree, I wish you'd tell me why.



    The integers and the rationals are isomorphic as sets, but not as ordered sets or topological spaces. Cardinality is a very weak metric for trying to say anything meaningful. There are some very wild countable sets.

    I don't see how the rational numbers have anything to do with this. When you say "countability preserves computability," I would challenge you to provide either a formal proof or a reference. It doesn't seem particularly meaningful to me; and to the extend it's meaningful, it's wrong.

    But like I say I'm just really trying to understand why you seem to be disagreeing with things I think we fundamentally agree on. Does that make sense?


    Yes yes it MAY WELL be capable of that, and SOMEDAY physicists may figure it out. But RIGHT NOW, contemporary physics has a thing called Planck time, which is the smallest interval of time we can meaningfully talk about. Below that there may be infinitesimal bits of time, and Zeno may live there, and so might Tinker Bell. As far as contemporary physics is concerned, any question or discussion about any interval of time smaller than the Planck time is meaningless.

    So you are AGREEING with me that we need NEW PHYSICS in order to get a better theory of infinitary computation that explains some of these mysteries.

    I think we're really in agreement on this point. We need new physics to get a new theory of computation that goes beyond Turing machines. This is something I do believe. We have to figure out how the universe might do an infinitary computation. Our current physics does not afford us such a theory.

    No reason at all, except for what contemporary physics says about the subject. Past that, one is engaging is scientific speculation.

    That's right. But our physics has no answers. Remember that the mathematical real numbers are not the same as the physical universe. We don't know if there are dimensionless points. We don't know anything below the Planck length and the Planck time. There could be dancing sprites down there, or Leibniz's monads. Nobody knows. It's speculation. And speculation is good, but let's label it as such.

    Metaphysical speculation. Perfectly fine, when labeled as such. "Probably do not exist." How would you know? How would I know? These are things nobody knows. Speculation.

    Sure but now you're arguing that the universe is random and the order that we see around us is an illusion, like seeing hunters in the sky. You just said, if I understood you correctly, that the qualities and attributes we see in the objects around us, are artifacts of our minds and not inherent in the objects themselves. Did I understand you correctly?
     
    Last edited: Feb 8, 2018

Share This Page