Will machines become conscious someday?

Discussion in 'Intelligence & Machines' started by Magical Realist, Sep 19, 2012.

  1. ElectricFetus Sanity going, going, gone Valued Senior Member

    Messages:
    18,523
    And again I ask: is that vital to connotative thought? If hormones are really only good for say emotion, then fuck that! I don't want emotional machines, you probably don't either! Is emotion important to general intelligence? probable not, digital machines have been programmed that can solve many specific problems, uprating that to general problems is not likely depend on emotion either. Back to connotative thought: is that dependent on emotion? Even catatonics claimed to have had some consciousness when frozen, and that about as devoid of emotion as you can get, lacking the desire even to move! A SAI may have desires though, hopefully ones we specifically program it with, like the desire to please humans and obey, might translate that as "emotion".

    Neuromorphic chips being analogy are likely to be inconsistent, they are "fuzzy" logically, and if we wanted to modify their pattern like a hormone would in humans we certainly could implement that in hardware with ease. Although why anyone would want to program PMS into a machine is beyond me!
     
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    If you asked that before, I missed it. Not sure exactly what "connotative thought" refers to but certainly it is not consciousness.

    I think this is answering/ responsive to/ your question:

    Certainly machines can respond and make wise decisions based on facts they have. In fact, again a recent study showed that on average they can use the available data better than the average of doctors to diagnose correctly illness and even recommend both the additional tests needed to discriminate between two possible illnesses AND when that is done, prescribe treatment plan / drugs etc. again with better results than the MD´s average! There is real hope that the growing medical cost with the aging population can at least be slowed if not actually decreased by machines replacing MDs.

    Many people will resist this, but few people can evaluate their doctor´s real skills.* I suspect the charteristic that earn the MD a high rating in fact tend to make you chose one with less than the average real medical skills and knowledge. – It is sort of like electing the US president. We tend to think the one with the more impressive office (or politician´s 30 second TV ads) or the one you think is a “nice guy (or gal)” is the best.

    I did have a one-shot test for new dentist: I asked if he knew what caused the two pure white spot on the tips of my top front incisors. Correct answer is that when they were forming I drank naturally high fluorine concentration water for a few months. If he did not know, I did not go back and once left immediately. (I don´t mind this cosmetic blemish – I have very hard teeth with no cavities, but now that I am old, four or five teeth have small corner chips missing.)

    * I had an MD dad, and have professionally worked part time with JHU doctors extensively, (~two decades) and thus have considerable medical knowledge in a few limited fields but still cannot tell a really good MD from an average one. I can usually identify one who is well below average in skills and jerks with a brief discussion of some medical ideas. I would prefer a well certified machine over the typical flesh and blood MD, diagnosing my problems except when the important test has a photograph as its output (e.g. X-ray or tomography, mainly). No machine can yet come close to practiced human vision for noting important details in a 2 D display.
     
    Last edited by a moderator: Feb 26, 2013
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. ElectricFetus Sanity going, going, gone Valued Senior Member

    Messages:
    18,523
    yeah was error check fucking up "cognitive". Again the question of consciousness is irrelevant because it impossible to answer definitively, thus the question should be "can we make a machine with general intelligence, a machine that can solve ANY problem presented to it as well as a human can or better. That easy to test for, the results are conclusive and irrefutable. Consciousness, a soul, these will always remain intangible, impossible to verify.
     
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    We agree:
     
  8. Write4U Valued Senior Member

    Messages:
    18,577
    question: If we could duplicate a brain atom by atom could it theoretically be conscious? If we gave it sensory information to the mirror neural network (say 14 years of schooling), would it become conscious?
     
  9. pitrodautility Registered Member

    Messages:
    3
    Although various types of hydraulic pumps have some common components, they still have much difference according to different modes of operation. For example, screw pumps are typically made up of a pair of intermeshed screws that rotate together inside of a metal housing.

    Thanks.
     
  10. kwhilborn Banned Banned

    Messages:
    2,088
    @ Write4U,

    No. Obvious reasons would be that you would need to create an entire animal not only for sensory input, but for survival.

    Still...

    Even still we often have perfectly intact brains (atoms are perfect) that are inside dead animals. Your creation would need to be alive. Perhaps in a perfect atom by atom structure your animal could simply get its heart moving via shock or massage, but it would still be dead upon creation.

    See
    Frankenstein
     
  11. someguy1 Registered Senior Member

    Messages:
    727
    The problem is that I can't even show myself another human that I can prove is conscious. Consciousness is a very difficult subject.
     
  12. Pithikos Registered Member

    Messages:
    74
    Can you provide a link to that study. Sounds like something interesting to read

    Please Register or Log in to view the hidden image!

     
  13. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    For one my age, I have a very good memory with millions of interrelated facts stored in it, so I seldom google and am not good at that, but your question seemed to be "just made for a Google search" even if with low skill. I have some medical problems and get Emails from several sources that review the latest developments. - I read one that said computers beat doctors especially in spotting rare diseases, instead of mis-diagnosing them as common disease. I have no idea what was the source of that was now.

    Here my first 2012 google hit: http://www.nytimes.com/2012/12/04/h...te-diagnostic-lapses.html?pagewanted=all&_r=0

    It mentions several computer programs that are in use. I suggest you search for more. I chose this hit as it is from about when I read the article, but not my article.
     
  14. Write4U Valued Senior Member

    Messages:
    18,577
    question: if someone is brain dead, but his body is kept functioning by artificial means, is that person dead or alive?
     
  15. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    Either way you want to define it. There is no absolute standard for what is living or dead. The more interesting question is when can the widow collect on the life insurance policy?
     
  16. spandrel Registered Member

    Messages:
    35
    The title of this topic is 'Will machines become conscious'. Algorithmic systems will always beat humans on some tasks but it's pointless to cite them as a step on the march to consciousness. Billions of people know nothing about the special subjects at which these systems excel, but we wouldn't deny that they are conscious and the computer program is not. So far no machine has passed the Turing test. Imagine a machine which kept bothering its human friend to tell it one way or the other whether it had passed the test. It wants to know if it's conscious or not. When do you stop the Turing test and make a decision? The test itself becomes an example of Turing's Halting Problem. The concept of self-reference, embodied in the Halting Problem, is fundamental to consciousness. We have no problem at all in saying "Oh, it's a paradox", an algorithmic machine can never do this. We can prove that there are some things a machine can never prove, including that. Weird.

    Is the brain algorithmic? The neocortex has bundles of neurons arranged in layers which have loads of lateral connections but not so many vertical connections to the next layer. This architecture is mimicked in artificial neural nets. There are six layers in these structures, not five or seven. Artificial networks don't even have six, as far as I know. These layers represent abstractions of information, so there are only six possible steps of abstraction. We can assume that nature, being efficient, has tried five and seven and settled on six. This alone is amazing! Only six steps to pull relevant patterns from any possible input.

    But this is algorithmic so is not consciousness. In fact consciousness has nothing to do with it. People who can achieve a deep level of meditation have no brain activity at all, as shown by fMRI scans. This bothers the hell out of neuroscientists, which is why they tend to ignore the problem. So something else is going on, a seventh level maybe, unconnected in a physical sense. Quantum shit? Does not compute.

    Recommended reading:
    Daniel Dennett - 'The Mind's I' explores the problems of consciousness.
    Roger Penrose - 'The Emperor's New Mind' explains why algorithmic machines cannot be conscious.
    Iain M. Banks - any of the Culture novels
    Alastair Reynolds explores whether a record of a brain is conscious or not in his 'Revelation Space' etc. novels.
     
  17. KilljoyKlown Whatever Valued Senior Member

    Messages:
    6,493
    This field is still advancing at break neck speed. I'm not sure anyone can say with any certainty what will be possible or not possible in the future. But it's an interesting work in progress, don't you think?
     
  18. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    Welcome to sciforums.
    It has been proven that a three layer "connection machine" (normally falsely called a "neural network")* can do any task the one with more layers can. (What is most important is the number of input, intermediate and outputs units, not layers, for the task.)

    Also nature, has achieve almost unbelievable things, but not always efficiently as it could be. Often it is trapped by the design the new creature evolved from. To give just two, well known examples:
    (1) human eyes work very well and are at essentially perfectly component matched. I.e. the entrance aperture of any optical system sets the resolution limit and the density of the rods in the fovea is just correct - more per unit area would not improve the max resolution and fewer would make it worse, but the retina is built "backwards." I.e. the last thing the light comes to is the photosensitive cells. The light must pass thru the network of nerves radiating out form the "blind spot" and all the retinal blood vessels. This is due to it being a part of the brain that prior to birth migrated to be part of the eye ball. The octopus has the correct design - photo sensitive cells are first with signal collection nerves and blood vessels behind that photo receptor layer.
    (2) The long necked giraffe evolved from a nearly no neck creature. One of the facial nerves in that ancestral "neck-less creature" which of course carried information form "face touch" etc. to the brain happened to pass below a bone (sort of a collar bone) and still does in the giraffe. I.e. if the giraffe's face brushes against a limb, the sensory signal travels all the way down the long neck, passes under the bone, and then back all the way up that long neck to enter the brain! - Not a very efficient design, I would say.

    * The only justification for this name is the Dr. Hebb (an MD) was one of the first to design, based on his understanding of how neurons learn, a learning rule for connection machines that worked well. Much better one are known known.
     
    Last edited by a moderator: Sep 14, 2013
  19. Mazulu Banned Banned

    Messages:
    3,090
    I'm still waiting for an intelligent robotic machine that can do dishes and laundry, mow the lawn.
     
  20. KilljoyKlown Whatever Valued Senior Member

    Messages:
    6,493
    Oh! Your not looking for smart machines then.

    Please Register or Log in to view the hidden image!

     
  21. Mazulu Banned Banned

    Messages:
    3,090
    What useful purpose would a smart machine serve?...other than to take my job. I'll take a few dumb machines.
     
  22. libiya Registered Member

    Messages:
    2
    Do you think so? Why or why not? And how would the invention of conscious machines by man influence the understanding of his own consciousness?

    Yes. Because technology is likely to increase, and Scientists are trying to make these types of machines. Man would simply understand that his awareness is not special because it arises in a different way or has different properties.

    No, and that wouldn't be a good idea because of neuroscience discoveries, the mind would literary still think it had a body. There are a lot of complications, it would be better to reverse engineer it instead.

    A smart machine would be another citizen of society, most likely.
     
  23. eugene381 Registered Member

    Messages:
    19
    Arguably, humans are machines, so conscious machines have existed as long as consciousness has existed.
     

Share This Page