I was thinking about Alan Turing ans some other functionalists the other day and I came up with a related question.
Even though the technology may not be present currently, do you think it is possible for a computer to correctly emulate the human mind?
One must remember that along with the memory and cognition of the human mind, what avout the consciousness and mental structure of the machine mind?

while reading your post, a small insight i just had was, while we as humans would want just the thinking part, to truly get a human level intelligence out of machines, wouldnt we also need the automatic parts as well, the regulatory, the subconcious, etc, the autonomic (i think correct part) nervous system, and the recognition that it is a mind. Wouldnt the best way of doing this be to model its development the same way you would a childs? make computer synapses that would start out with just the few basic connections to regulate its growth, as in a child, then have it "grow" into itself? For instance, have the synapses make new connections much as ours do to uniquely store the information that it locates?

that was my first impression after reading your post,
tootles!

No, a computer cannot do anything complex. A computer is a simple linear array of registers with a fetch and decode machine. It has no sensors, no brain, no motors.
I am not saying this lightly, I started working on AI in 1958 and gave it up in 1975. That was when I realized that even if you had a program to emulate a brain, it would not do anything and would be worthless.
We should get rid of the idea.

hmm a computer is only as smart as the person who inputs the data, i think that in the future yes the cpmputer will be complex enough that it can imitate the thought processes of the human brain, i dont think that i will see it in my lifetime but i think in generations to come computers and humans will become absolutely dependent on one another, look at the facts what is there that cant be done over the internet, in every aspect you can some how link it back to a very simple binary strand of data, our brain has only a few things over the computer it has:
a) the ability to reason

b)the ability to feel affection
but i think that the next evolution of man kind will be dependent on the simple processes of computes so brains will do more complex things

i hit tab enter, hence the two posts sorry

Originally posted by hlreed
No, a computer cannot do anything complex. A computer is a simple linear array of registers with a fetch and decode machine. It has no sensors, no brain, no motors.
I am not saying this lightly, I started working on AI in 1958 and gave it up in 1975. That was when I realized that even if you had a program to emulate a brain, it would not do anything and would be worthless.
We should get rid of the idea.

thats using linear computing though, what about quantum computing? their non linear arent they?

or analog computing?

If you all think about it, our brain is just a complex computer in itself, so isn't asking this begging the question: can we build a computer that can emulate our natural computer? Of course.

No it can't be that stupid!
:D

Originally posted by Fafnir665
thats using linear computing though, what about quantum computing? their non linear arent they?

my thoughts exactly - i'm on another thread with someone about quantum computing in relation to the theory of parallel universes put forth by David Deutsch, and I agree that a linear system can't emulate a nonlinear system (which I believe the mind to be) - but what about quantum computation?

Isn't it interesting that the binary system used by classical computers is only analagous to reality on a certain level? On the quantum level, all of those rules break down. Something can be more than one thing at once at the quantum level. Perhaps the "mind" follows different rules than the "brain" in the same way that different rules apply at the quantum level for matter. So - would a quantum computer be able to emulate the not-so-black-and-white states of consciousness, emotion, etc. (as distinct from linear reason)? I'm especially interested to hear the opinion of the guy whose done work in AI.

depends on what you mean by emulate a human mind. do you want it to work in the same exact way? or do you just want similar outputs and behaviour?

computers won't work the exact same way. think about memory. a computer can had a 1-1 perfect representation of all it went through in memory. that is very unhuman like. even thought it's not like a human in function you could still have it be as 'intelligent' as a person.

i think we should recognise that human intelligence is not the only kind of intelligence there can be and that any intelligent machines we create in the future may or maynot be just like us.

Oh, I agree with you. Besides, what would be the POINT of creating something EXACTLY like a human? What would be the value in creating something with all the same flaws? To be able to create any kind of intelligence would be remarkable - a carbon copy of human intelligence would be kind of . . . useless.

[quote]
computers won't work the exact same way. think about memory. a computer can had a 1-1 perfect representation of all it went through in memory. that is very unhuman like. even thought it's not like a human in function you could still have it be as 'intelligent' as a person.
[/qoute]

Computers can be programmed to be imperfect.. Even indeterminate, impossible to predict.. Computers can be just as frail as humans..

Blindman is perfecly right , a computer can be programmed to forget ...

and I do think computers could replicate a human mind it's just that you'd probably have to simulate every atom (actually every subatomic particle) in the HEAD and the way they interact to get a near perfect replica ... unfortunately it would be almost impossible to make it work in "real time" because of the way computers are built ...

simulating only our neural network might work faster but it would still require a huge amount of computing power and you would have to replace some of the "randomness" and the effect of chemical substances and many other things that make up a human mind...

and

No, a computer cannot do anything complex. A computer is a simple linear array of registers with a fetch and decode machine. It has no sensors, no brain, no motors.

how can someone that has been working in this field for so many years bring such a silly argument ?!? ... if you want to replicate a human mind you would have to also at least simulate all inputs and analyse the outputs ... and although I usually respect older people I must remind you that you gave up on the ideea when computers were still very very useless and very slow ...

however it is true that simulating a human mind is completly useless and a waste of time ...
studying the mind and creating human like behaviour or even superior intelligence is a completly different thing and must be pursued

I think it is rather simple why we cant replicate a human mind...

1st: There is still alot of the brain that is undiscovered.
2nd: There is problems with the input... humans have 5 senses that gathers data, a humanoid robot has 3 at most...

Originally posted by Jez
I think it is rather simple why we cant replicate a human mind...


sorry, but neither of theses count as simple, they may be easy to write down, but even three words can carry a lot of complexity, such as "i love you", but thats another point...


Originally posted by Jez
1st: There is still alot of the brain that is undiscovered.

i thinkw e realized this... at least i do, and thats just a reason why we can't do it now...


Originally posted by Jez
2nd: There is problems with the input... humans have 5 senses that gathers data, a humanoid robot has 3 at most...

Thats not really a good reason, yea we have 5 senses, and theres only three available for the robot, but who says we can't create different sensors? The statement is false anyways, robots have ears, eyes, dispersed pressure sensors (touch), olfalctory (smell) though it is in its infancy at the moment, and taste, again rudimentary, and "sense" that humans have can be broken down to some kind of chemical reaction, so its not correct to say that a robot can't posses the few sense we have, plus more.

1.There are ways to find out how the brain works and people are working on it , it's only a matter of time before we will be able to fully understand it.

2.Robots can have virtually thousands of senses and can gather alot more data then a human ... altough if you want to create a human mind you don't need it.

Ya, the question is not whether someone can build one in the next year, its if its possible; and that answer is obvious.

Originally posted by Frencheneesz
Ya, the question is not whether someone can build one in the next year, its if its possible; and that answer is obvious.

i dont see any of that in the original post, what are you reading?

Fafnir665:

Well, I was refering to this question:

"Even though the technology may not be present currently, do you think it is possible for a computer to correctly emulate the human mind? "

Do you see it now?

yes, i see it now, your still reading it wrong, it asks if it may be possible, it has no relation to a time period

my point exactly.... ?

Originally posted by Frencheneesz
my point exactly.... ?

oh yea, i read it wrong.... *puts away the crack pipe*

ha lol. Ya get some weed instead, then you can't even type.

Question. how close are we (if we wanted to) to take a human brain (adult and living) take it out of the body and keep it functioning artifically?

For a long time I refused to accept that there would be any economic benefit to producing a human simulation. The research investment would make the space program look like a walk in the park..

But recently it has been pointed out to me, that there is a very good source of money that will drive the development of the human simulation. Medical research. Every human is interested in the development of medicine. There are many software models of humans already in the pipeline. Just recently I had the opportunity to play with a 3D simulation of a hand. I had a force feedback simulation of a scalpel, I could cut, poke, pull. I could feel the bone. Geez I was using the needle just to see if I could get it in between the joints. This was just the demo and it was at least 4years old.

I think that the main point is. When is the human abstraction considered to be human?

Also it is impossible to simulate anything 100%. The only way is to simulate a single thing is to simulate the entire universe including the simulation its self and that is totally impossible.

Dtroit and all,


do you think it is possible for a computer to correctly emulate the human mind? Emulate, enhance, and replace completely.

Skull Size

The growth of the human brain is limited to the size of the skull. It is going to take many thousands of years before biological evolution changes that scenario. With genetic engineering we could probably shorten that period to hundreds of years.

Computing Power.

However, computer technology is only a few years away from having the computing capacity to equal the human brain; we just need the software to catch up. But software development will progress much faster by using the newer faster hardware technology. This results in an exponential feedback loop. If Moore’s law continues to hold then once we achieve human level processing power in a single system then 12 to 18 months later we will have doubled that capability, then quadrupled etc.

Neurons.

The human brain appears apparently complex, but it is essentially a vast network of interconnected simplistic neurons. We could say the Internet is complex in the same measure but again it is just a vast network of interconnecting nodes. The difference is that the internet is connected by thousands of nodes but the brain has some 100 billion nodes very closely located.

Brain/Computer Analogy.

We should stop comparing the brain to a computer. The idea evoked is that of the desktop machine that we all know. But the brain doesn’t have a single CPU executing one instruction at a time. It is more appropriate to think of each neuron as a small microprocessor and with 100 billion of them all operating in parallel and communicating with each other. An analogy to the internet is closer.

Digital versus Analogue.

Each neuron fires at around 200 times per second. That is not an analogue process. Something operating digitally and fast enough appears analogue. Consider a movie; it is separate frames being moved rapidly. Think of a music CD – that is totally digital.

Hlread,


No, a computer cannot do anything complex. A computer is a simple linear array of registers with a fetch and decode machine. It has no sensors, no brain, no motors.And a neuron is similarly simple. Now connect 100 billion of them together, and you can produce what many would recognize as complexity.


I am not saying this lightly, I started working on AI in 1958 and gave it up in 1975.At that time there was immense, but unrealistic, hope for the future of the computer with AI, but those early proponents soon became tremendously disillusioned when they could not produce anything like AI. Their fatal mistake was to completely underestimate the massive computing power that would be required. Even now we will require the equivalent of some 10,000 2GHz PCs to achieve equivalent processing power of the human brain. That is a level of power almost unimaginable in the 1958-1975 timeframe.


That was when I realized that even if you had a program to emulate a brain, it would not do anything and would be worthless. In 1975 you would be quite right. But now I think you should update your perspective.


We should get rid of the idea.Then do so at your own peril. Many others around the world are very actively pursuing AI and with the much needed massive computing power on the near horizon AI will materialize much sooner than most expect. The continuing exponential increase in computing power will allow AI to appear and then supercede human intelligence very rapidly. Where does that leave humans? No longer the dominant intelligence on the planet.

I think what hlreed meant when he said it was useless was that an AI, by itself without external input, would be useless. I agree. The human brain seems to do VERY weird stuff when all its external input is taken away (sensory deprivation). However, computers have progressed to the point that thousands of types of input are available for the "mind" of the AI. As well as an almost endless amount of information available if it could read the internet. The hyperlink system was designed for the human mind because of the way it links similar types of information. I think perhaps this may be a great way to teach any AI. I think a largely missing input is touch. The human body has a HUGE amounts of pressure sensors all over its body. Most computers (read robots) have very few, if any pressure sensors.

-AntonK

Originally posted by Cris

We should stop comparing the brain to a computer. The idea evoked is that of the desktop machine that we all know. But the brain doesn’t have a single CPU executing one instruction at a time. It is more appropriate to think of each neuron as a small microprocessor and with 100 billion of them all operating in parallel and communicating with each other. An analogy to the internet is closer.



I agree. I think one of the most important factors to the human brain is randomness and degrees of randomness (if there is such a thing). What causes a human brain to continuously switch from one thought to another? –it’s not actually a “switch” but a gradual flowing shift of thought. Somehow, billions of (mostly) random outputs from microprocessors would be collected and produce a single thought. The direction or tendency of these outputs would be affected or swayed by the senses. An appropriate analogy is the brain working using the rules of evolution. Random mutation (random outputs with the potential of producing millions of different thoughts) combined with adaptation to the environment (the senses sway or give direction to achieve the most beneficial thought). I hope I made some sort of sense!

The data flow in anything with a brain is:
nature -> sensors -> brain -> motors -> nature
Note the loop. Sensors read nature. Motors modify what sensors see.
From this we deduce that brain mediates between sensors and motors. A brain is always between sensors and motors, meaning anything that has more sensors than motors must have a brain.
Every cell in your body has a brain. Every microbe has a brain. Every neuron has a brain.
Neurons are functions of many inputs and one output. To make parallel functions, you combine two input functions into a tree. An n input tree requires at least n/2 functions, each of which must be a microprocessor.
I have an algebra to build these. I have the hardware for the functions. I have them for sale.
There are several problems. The functions are low level and a lot are required. The cost is still too high. And nobody believes me.
The final problem is my brain may go before this gets done.

No, a computer cannot do anything complex. A computer is a simple linear array of registers with a fetch and decode machine. It has no sensors, no brain, no motors.
I am not saying this lightly, I started working on AI in 1958 and gave it up in 1975. That was when I realized that even if you had a program to emulate a brain, it would not do anything and would be worthless.
We should get rid of the idea.We have not failed, we've only found 10,000 ways that do not work. ;)

Of course there will be AI that can emulate a human brain. It's inevitable.

hlreed,


There are several problems. The functions are low level and a lot are required. The cost is still too high. And nobody believes me.
The final problem is my brain may go before this gets done.Ha ha, don't worry, I believe you. And yup I'm concerned that I won't be able to upload before my brain deteriorates as well.

But I'm hoping for a lot from Intel and their Itanium family especially as they proceed to multi-core chips. So a 32 core Itanium loosely coupled should simulate the activity of some 500 million neurons. That leads to around 200 chips linked via a high bandwidth SAN. That’s about the year 2010. And if we say 4 chips to a blade and 25 blades to a standard rack then that indicates we’d need two 8 ft standard racks. That doesn’t sound so bad.

Here is the basic problem with AI. Go back to the data flow.
nature -> sensors -> brain -> motors -> nature

Outside we build sensors -> brain -> motors. Nature is nature.

Inside the computer we must simulate nature as well. So the whole structure of the simulation is fabricated.

It is a simulation problem, not a problem of speed of computation. The problem is fundamental and cannot be fixed by tweaking.

The simulation must be taken out into the world for its inputs.

The same applies to artificial life as currently implimented.

Take them both out of the computer and we come up with the same animachines for both. Then you can remove the artificial attribute.

hlreed,

I dislike the term AI. I don't think intelligence can be artificial. This implies that it is not real. But if something is intelligent then it is intelligent. What is really meant by AI is non-human intelligence.

The name I gave to this forum 'Intelligence and Machines' was a deliberate attempt to avoid the 'artificial' stigma with this arena. So MI (Machine Intelligence) or NHI (Non-Human Intelligence) would be preferable to AI.

But you are implying that intelligence cannot exist without an external interface. I suspect that is quite untrue. Providing the intelligence has some data to work with then I'm sure it could exist. How we would know it was intelligent or how it could obtain useful data is quite a different issue? Imagine a human who has had all sensory input and communications methods severed and is in a true state of sensory deprivation. Ignoring the potential for madness, the intelligence can still process thoughts, make deductions, experience emotions, etc. What is removed is the ability to obtain new external data or to communicate any decisions.

But having lots of external stimuli is pretty useless until the mechanism to support the intelligence exists. Creating the framework of an intelligent machine should be possible, e.g. a biological equivalent would be a brain with 200 billion neurons but none of them have formed any connections or neural networks. Now turn on the external sensors and allow the neurons, or the simulated machine neurons to form networks.

The most interesting milestone will be self-awareness and I am assuming this will be achieved when a minimum threshold level of neural networks is reached.

Imagine a human who has had all sensory input and communications methods severed and is in a true state of sensory deprivation. Ignoring the potential for madness, the intelligence can still process thoughts, make deductions, experience emotions, etc. What is removed is the ability to obtain new external data or to communicate any decisions.

... sensory depravation will only cause emotions and so on if you apply it to a rather "mature" brain ... if you were to do this to a small child his brain would probably die slowly because it is not needed - actually I think this would happen even to an older person , it just might take a little longer
and yes , A.I. is not a very usefull word for describing what we're looking for and we must stop thinking of computers as the little box on our desktop - I actually have a ... hmm "project" ... but it's too far fetched ... so I'm working on another project that will help me to work on every other project I want - you could say it's a lab ... or something

I dislike the term AI. I don't think intelligence can be artificial.Why not? If intelligence can arise naturally, I see no reason it can't arise artificially.

This implies that it is not real.No it doesn't. It implies it was created by man.

I don't see why you have a problem with the term. It's fairly accurate.

Ya, artificial means it was fabricated by human means, it doesn't mean it is imaginary. You have artificial fur (clothes), yet by no means is it not real.

But what does one mean when they say intelligence? If you don't believe in a soul, the brain is just a very complex and non-direct cause and effect system. Just because we don't understand it doesn't mean its random. The brain is VERY complex, but each neuron is simple enough to understand individually. The complex part is how to organize those neurons so that they do something useful.

Alpha,


If intelligence can arise naturally, I see no reason it can't arise artificially.
It implies it was created by man.
I don't see why you have a problem with the term. It's fairly accurate.Yes I agree. I’ll update my erroneous interpretation of the term.

Many believe computers will surpass the human brain in complex processing ability within 40 years.

How can we resist designing computers that mimic the brain? Imagine a machine that not only mimics, but surpasses the human brain in many ways. Will this machine claim to be alive? Will it begin to ponder it's own mind as it learns, and in doing so, believe it may have a soul? When it shows us that it is every bit as alive as we are, what will we think? Some of us will see this as proof that the soul doesn't exist. Others will believe the machine somehow captured a soul from the world beyond or gave birth to one. Whatever, we'll have quite an enigma on our hands should a machine that surpasses human intelligence insist in a very convincing way that it is perhaps more alive than we are.

I found this site:
The Kurzweil Hypothesis: http://www.zianet.com/khacker/kurz1.htm

damn souls. Thats the only idea that I can't attribute to science. Why are we conscious? Why aren't our particles just working in mechanism to produce what is here? Why can I see and live my life? It doesn't make sense, souls don't make sense. Do you believe animals are conscious like we are? at some level creatures cannot be conscoius, a rock for example. So animals might have consciousness, but I'm only sure humans have consciousness, because humans are like me, and I know I'm conscious..

Originally posted by Fafnir665

Thats not really a good reason, yea we have 5 senses, and theres only three available for the robot, but who says we can't create different sensors? The statement is false anyways, robots have ears, eyes, dispersed pressure sensors (touch), olfalctory (smell) though it is in its infancy at the moment, and taste, again rudimentary, and "sense" that humans have can be broken down to some kind of chemical reaction, so its not correct to say that a robot can't posses the few sense we have, plus more.

I can see in infrared,UV, radio wavelengths, I can hear in sub bass and ultra sound, there's millimetric waves for seeing through clothing and I can tell what voltage is going through a piece of wire, and all this is because humans have hands.

Alot of people forget that a brain has a body. Take 1 baby, extract it's brain with eye's attatched, splay out it's touch sensors into a square and put it in a room for the rest of it's life.
It sounds sick to me as well but thats what some of you are trying to do with A.I.
It needs a body connected to the brain, you can only design a brain when you know what the body will look like and what sense it will have.
We are not designing a replacement for humans we are attempting to design something that resembles a human, it won't be human as we know the term human.

Anyone up for designing F.R.E.D the electronic ape. I got some basic drawings.

Originally posted by Frencheneesz
damn souls. Thats the only idea that I can't attribute to science. Why are we conscious? Why aren't our particles just working in mechanism to produce what is here? Why can I see and live my life? It doesn't make sense, souls don't make sense. Do you believe animals are conscious like we are? at some level creatures cannot be conscoius, a rock for example. So animals might have consciousness, but I'm only sure humans have consciousness, because humans are like me, and I know I'm conscious..

you think that's mind boggling question now, add this occurance in there-
drat, looks like the article has been removed from new scientist. About a year or so ago, they reported on a microchip that was designed to evolve over time to produce a design that would create a designed effect. For one experiment, the chip was given the task of generating an ocilating frequency. which it did. Thing was, when the scientists went to look at what the final design the chip had come up with, they found something completely different than they were expecting.

THe chip had developed a radio antenna of sorts, and was picking up the EM radiation from a server in the next room. it was then broadcasting this signal as it's oscilating wave. Given that the scientists had not provided the chip with any specific method to pick up external radio signals, how in the hell did it notice the EM waves enough to evolve a method for picking up and broadcasting those waves????

here it is, I found it:
http://www.newscientist.com/news/news.jsp?id=ns99992732

If we did create a computer to emulate the human mind, would it be as confused and crazy as we are? Would the massive amounts of parallel computing lead to the same "fuzzy" thinking that we all possess? I assume everyone here has read "Godel, Escher, Bach" by Hofstadter? I hope I spelled the name correctly.

Wind-river :

I think that oscilating circuit was just an accidental radio reciever. The EM transmition in the other room is picked up by any metal surface in rance, so the circuit picked up the transmition just like any chair or fork would. However since the circuit is a circuit, it was altered by the EM and gave a different output. Thats what it looks like to me.

Oh, I expect there will be many different types of mind that emerge out of AI research-
there could be emulations of the human mind, running at human speed, or faster than human speed, with the fuzzy thinking that allows us to make wild-assed guesses and to create art;
the minds could be similar to humans, but with the fuzzyness and poor memory capacity improved or capable of focus;
there could be minds tens, hundred, millions of times as complex as human minds;
artificial minds developed from insect templates, mice, dogs, chimps, autistic humans, geniuses;
developed by programming from the top down, or emerging from the bottom up;
developed from automatic feedback systems and internet search engines.
Minds that are enslaved to their programming andminds that are free to enslave.
The possibilities are endless...
we have begun to explore some of these possibilities in our imagineered scenario
Orion's Arm, but there are bound to be many more possibilities...
including the (frankly remote) possibility that AI is not possible at all.

__________________
SF worldbuilding at
http://www.orionsarm.com/main.html

Computer emulate human mind? Why not? I think it will require many though:
-patience-> on making every single response possible when the computer is given certain action.
-hadware-> too many to store= TBs
Else? I haven't found out for sure....