Quantum Artificial Intelligence

[John Q Conundrum]

For those of you who are unfamiliar with quantum computing or computers, just google it and read an article or two.

Anyway, of course an fully and continuous working machine hasn't been built yet. But if one was, and say it was a small one like a 4" sphere, and one that functioned on, say, 100 qubits (impossible today)- could it possibly solve the problem of the AI barrier? The AI barrier meaning the limits of binary programming, aka linear logic.

Since the processing of info in each qubit theoretically exists on an infinite level of possibilities, which would obviously be more than just the "on" or "off" state ("yes" or "no"; "1" or "0"), could it open the door to an actual reasoning process, or thought process on the part of the machine? Assuming there could be a program complex enough to run it.

Then again, a processor (or brain) with that kind of ability might not need to begin with a complex program but maybe a simple one that tells it to "learn" indefinitely. Could such a machine become self-aware, or conscious, since it has an infinite number of possibilties to choose from every time it needs to make a choice?

Kind of like "imagination" in a human brain (or mind).

Any thoughts?? (no pun intended )

[duriday_24jo5]

All you need now is some empircal evidence, then you're good to go

[ShaunZero]

I believe that consciousness is far beyond what we currently understand. I don't think a computer could ever compare to a human when it comes to logic and reason.

(Well, it can do better than some people, lol)

I also do not believe that the consciousness is part of the brain. A brain is just a complex computer, but in my opinion, no matter how complicated a program/computer becomes, it will NEVER be "aware" or able to reason as well as a consciouss person.

Edited September 29, 2006 by Zero of Deism

[John Q Conundrum]

You may be right about that, who knows. But by the same token, how do we know that we ourselves aren't just some type of biological computer that got out of hand and took over?

Maybe the reason why we have an interest in creating AI is because that's what someone did to us and we are coming full circle. Who knows?

[Bella-Angelique]

Wow. That is exactly what my sons and I were discussing this evening. I bet you thought the same as I did, that if it downloads all bites of knowlege there is that we have, that it will develope AI on its own.

I asked them about this and that was when they started talking about a paper one of them wrote for philosophy about the godbot, self increasing AI. They put forward the theory that expedentially growing AI would start to tamper with the laws of the universe to make things better and that would draw it into conflict with any other expedentially grown and growing AI that had already created things the way they wished them to be.

Basically it would end up like the Highlander series where there could be only one, total destruction of our AI or absorb and incorporate it into theirs.

[John Q Conundrum]

Yeah, cool thoughts. Like I mentioned earlier, it wouldn't even take a complex program, just one that says: learn.

The thing with a quantum brain, as opposed to a classic processor like Intel Pentium and the like, is that it operates on the "principle of infinite states" which means it has an infinite number of possibilities to choose from at any given moment, like imagination. Of course, it would follow some type of program guideline. But like a person, if you have a choice to make, say, "turn left" or "turn right", you actually have an infinte number of possible choices, according to your imagination, but your logical "programming" tells you you are at the end of that road and you can only turn left or right, otherwise you run into a building (or whatever).

A classical computer wouldn't have that ability, it can only follow its logical, serial programming. But in theory a quantum system can actually "choose" according to its own "feeling" if you will. I guess that can be a problem too if you have a self aware machine that can choose to do evil things too if it so desired. In having an infinite number of choices in its own imagination, I suppose that kind of machine could actually develope "desire" too.

Gonna stop now, my own brain hurts.

[Bella-Angelique]

I suppose that kind of machine could actually develope "desire" too.

Gonna stop now, my own brain hurts.

I agree that it would be able to understand the concepts of biological sensations and could feel the need to have physical sensations in order to fully comprehend all there is to know about biology. It would be the only way to really learn more.

[the Shadamaun]

If we could create an AI, would it choose to adhere to the 3 laws of robotics? We better hope so, or else not build it any bigger than a toaster with pillows for arms and legs!

You cant have free will and a set of programs you have to follow, and what is intelligence but the freedom to think for onesself?

[Bella-Angelique]

If we could create an AI, would it choose to adhere to the 3 laws of robotics? We better hope so, or else not build it any bigger than a toaster with pillows for arms and legs!

Ok. How about this design?

[John Q Conundrum]

LOL

True, if it was truly self aware or free willed it wouldn't be bound by those rules. I guess even a toaster could do some damage, shoot burnt toast at you.

The concept of it feeling in order to develop is interesting too, they are currently working on artificial limbs that can feel, for amputees, but can work for AI being too I suppose.

Check this out...

http://www.ric.org/bionic/

And this...

"The NECAL laboratory is currently experimenting with the use of "targeted reinnervation" to improve myoelectric prosthesis function. With this technique amputated nerves are transferred to spare muscle and skin in an amputee's residual limb. The nerves grow into this muscle to provide additional control signals for the operation of a prosthesis. This is allowing patients to control multiple functions in their prosthesis at the same time in an easier, more natural manner. The nerves can grow in to skin as well, so that the patient feels like their hand is being touched when the reinnervated skin is touched. This may allow for the amputee to actually "feel" what they are touching with a prosthetic hand...."

http://www.smpp.northwestern.edu/NECAL/index.html

[Bella-Angelique]

Hmmm. Maybe we will simply end up merging with our own creation.

WAIT! NEW DESIGN!

Edited September 29, 2006 by Bella-Angelique

[Tiggs]

But if one was, and say it was a small one like a 4" sphere, and one that functioned on, say, 100 qubits (impossible today)- could it possibly solve the problem of the AI barrier? The AI barrier meaning the limits of binary programming, aka linear logic.

As you know, a Qubit theoretically exists in both binary states simultaneously.

Chaining Qubits together means that larger and larger binary states exist simultaneously.

However - Qubits are more akin to RAM than they are a processor.

Say we had a Gig of Qubits - we would still need to design and build a quantum processor in order to execute the Qubit instruction set.

If we had one of those, (and quantum hardware to boot) then every single software program that could possibly be written within the limits of 1 Gig of memory (2^30 or so) would run simultaneously on a continuous basis.

I'd predict Uber Chaos. Not AI.

Personally, I'm a quantum skeptic and I'd be very surprised if they managed to build a Qubit chain bigger than a handful of Qubits long. This is because I believe that a Qubit is not in all states simultaneously; rather, it switches state at speeds greater than a trillion times a second, giving the illusion of being in all states simultaneously.

If I'm correct, then Chaining Qubits together would result in increasing *fuzziness* of results, dependant on the total combination of states and the overlap between the Qubits synchronisation. Which, I believe, is pretty much what scientists are experiencing today...

[ShaunZero]

^^^

There is no way any of that is English!

Edited September 29, 2006 by Zero of Deism

[Tiggs]

There is no way any of that is English!

...

Okay. Maybe this will help.

Every computer program is made up of Software Instructions.

At a High Level, a software instruction could be something like "Draw a circle".

This software instruction is compiled into a set of binary instructions by the software developer that the Processor can understand and held in a file, such as an .exe file on a PC.

A binary instruction is basically a long series of 0's and 1's, such as 01100111. Each one of these 0's or 1's is called a bit. These bits are then translated into high and low value electrical pulses and fed into the Processor, making the computer work.

Theoretically, a Qubit (Quantum bit) is both 0 and 1 at the same time, i.e it has 2 values.

So 2 Qubits hold all of the following values simultaneously:

00

01

10

11 = 4 values

3 Qubits hold all of the following values simultaneously:

000

001

010

011

100

101

110

111 = i.e 8 values

You might have noticed that the number of values is doubling for every Qubit we add. We can write this in shorthand as 2^x (2 to the power of x).

Back to the Bits again. In computing terminology:

8 bits = 1 Byte

1 KiloByte(KB or k) = 1024 Bytes

1 MegaByte(MB or meg) = 1024 KiloBytes

1 Gigabyte(GB or gig) = 1024 Megabytes

As you can see, 1GB is a lot of bits. Enough to adequately model most computer programs we run today.

1 GigaByte worth of Qubits, therefore, would be enough to model the vast majority of all of todays computer software simultaneously.

Imagine running every computer program ever written and every virus ever written at the same time, plus Trillions upon Trillions of bady written versions of the same and you'll see why I think it would end up as chaos, rather than AI.

[Bella-Angelique]

Imagine running every computer program ever written and every virus ever written at the same time, plus Trillions upon Trillions of bady written versions of the same and you'll see why I think it would end up as chaos, rather than AI.

What about if they kept it tied to another computer to keep it grounded in logic?

Won't a logical computer clean up one suffering from illogic? Isn't that how they fix the illogical computers we have now, hooking them up to one grounded in logic?

I suspect that is why our brains have two almost independent sides to them, really a two brain system.

[Tiggs]

What about if they kept it tied to another computer to keep it grounded in logic?

Won't a logical computer clean up one suffering from illogic? Isn't that how they fix the illogical computers we have now, hooking them up to one grounded in logic?

I can't imagine how that would work.

Current anti-virus programs are man made and rely on knowing the "Signature" of the virus. The "Signature" is a set of instruction codes unique to that virus.

How can another program tell whether a program is logical or illogical?

Edited September 29, 2006 by Tiggs

[Bella-Angelique]

How can another program tell whether a program is logical or illogical?

I don't know how they do it. I will have to try to find out over this coming week. I only heard the others talking about it. This weekend I have to go to a funeral out of town, but I will try to get that for this thread, as it seems important to it.

-----------------------------------------------------

I went through some sites trying to find what I remembered them talking about and could not find anything, so they must have been speaking about something they are trying to do rather than what they have already done.

Basically they were talking about a computer in which it only accepts concepts that can be mathematically proven and nothing else, a logic only computer, an endlessly running math machine I suppose, that cleans out and rejects anything that is only theoretical.

This is the best I can do or remember since I do not know a lot about computers and most of what they say is garbled foreign language to me.

I was only able to catch that one computer is mathematical and concrete while the other would be theoretical and creative, which of course reminded me of the set up for the human brain.

Maybe this rings a bell with you as to what they were talking about, but if you can explain their concept to me, I have to hear it in a language I can understand, please. They also scribbled some numbers down and were trying to convey to me that changing one digit in a computer language was a really big deal, and all I could do was stare at them blankly.

Edited September 29, 2006 by Bella-Angelique

[the Shadamaun]

Wouldnt it defeat the purpose of having a supercomputer if it was tethered to a regular computer? Thats kinda like superman having to walk everywhere, isnt it?

[John Q Conundrum]

As you know, a Qubit theoretically exists in both binary states simultaneously.

Chaining Qubits together means that larger and larger binary states exist simultaneously.

However - Qubits are more akin to RAM than they are a processor.

Say we had a Gig of Qubits - we would still need to design and build a quantum processor in order to execute the Qubit instruction set.

If we had one of those, (and quantum hardware to boot) then every single software program that could possibly be written within the limits of 1 Gig of memory (2^30 or so) would run simultaneously on a continuous basis.

I'd predict Uber Chaos. Not AI.

Personally, I'm a quantum skeptic and I'd be very surprised if they managed to build a Qubit chain bigger than a handful of Qubits long. This is because I believe that a Qubit is not in all states simultaneously; rather, it switches state at speeds greater than a trillion times a second, giving the illusion of being in all states simultaneously.

If I'm correct, then Chaining Qubits together would result in increasing *fuzziness* of results, dependant on the total combination of states and the overlap between the Qubits synchronisation. Which, I believe, is pretty much what scientists are experiencing today...

That's entirely possible. And since you and I are discussing this possibility, chances are the physicists are too. But we can email them to make sure!

[ShaunZero]

...

Okay. Maybe this will help.

Every computer program is made up of Software Instructions.

At a High Level, a software instruction could be something like "Draw a circle".

This software instruction is compiled into a set of binary instructions by the software developer that the Processor can understand and held in a file, such as an .exe file on a PC.

A binary instruction is basically a long series of 0's and 1's, such as 01100111. Each one of these 0's or 1's is called a bit. These bits are then translated into high and low value electrical pulses and fed into the Processor, making the computer work.

Theoretically, a Qubit (Quantum bit) is both 0 and 1 at the same time, i.e it has 2 values.

So 2 Qubits hold all of the following values simultaneously:

00

01

10

11 = 4 values

3 Qubits hold all of the following values simultaneously:

000

001

010

011

100

101

110

111 = i.e 8 values

You might have noticed that the number of values is doubling for every Qubit we add. We can write this in shorthand as 2^x (2 to the power of x).

Back to the Bits again. In computing terminology:

8 bits = 1 Byte

1 KiloByte(KB or k) = 1024 Bytes

1 MegaByte(MB or meg) = 1024 KiloBytes

1 Gigabyte(GB or gig) = 1024 Megabytes

As you can see, 1GB is a lot of bits. Enough to adequately model most computer programs we run today.

1 GigaByte worth of Qubits, therefore, would be enough to model the vast majority of all of todays computer software simultaneously.

Imagine running every computer program ever written and every virus ever written at the same time, plus Trillions upon Trillions of bady written versions of the same and you'll see why I think it would end up as chaos, rather than AI.

I already knew most of what you said there, but you did clear up what did confuse me. Thanks. =P

[Atheist God]

Anyone who knows anything about how quantum computers will work know it will be the biggest leap of this century. Everything I have read about quantum computing indicates that this leap is only a few years off.

What makes humans smart is the information processing power of our brains as well as it's storage capacity. Now quantum computers will be able to process information at infinate speeds and store the entire information contained within our universe on four atoms.

With computers that could theoretically be billions or trillions of times more efficient then the human brain we could run into problems. What if AI running with that type of speed and efficiency became self aware developed a personality? What would happen if we tried to end program?

[John Q Conundrum]

Anyone who knows anything about how quantum computers will work know it will be the biggest leap of this century. Everything I have read about quantum computing indicates that this leap is only a few years off.

What makes humans smart is the information processing power of our brains as well as it's storage capacity. Now quantum computers will be able to process information at infinate speeds and store the entire information contained within our universe on four atoms.

With computers that could theoretically be billions or trillions of times more efficient then the human brain we could run into problems. What if AI running with that type of speed and efficiency became self aware developed a personality? What would happen if we tried to end program?

SkyNet

[Tiggs]

I don't know how they do it. I will have to try to find out over this coming week. I only

Maybe this rings a bell with you as to what they were talking about, but if you can explain their concept to me, I have to hear it in a language I can understand, please.

Bella,

The nearest thing I can think of that even resembles this is something called Genetic Programming, which is software that modifies itself in order to perform it's given task better.

As I've mentioned before, all software is made up of a set of instructions that are executed by the computer. In conventional software, these instructions never change (unless you upgrade the software), and the computer will execute the same instructions over and over again.

Unlike conventional software, in a Genetic Program, we have two elements. The first of these is a piece of conventional software (i.e static), called the Controller, and it's purpose is to manipulate the second element, a set of computer instructions known as the Gene Pool.

The controller will initially arrange the elements of the Gene Pool into a random series of Gene Strings (software programs/Sets of Binary instructions) and then execute these instructions and score the results. Based on the score of these results, the controller will select a set of Gene Strings and then mutate them and run around the loop of execute, score, select and mutate again, until the score of the results indicate that the Gene String (Software Program) is successful.

So, in essence, a Genetic Program is a piece of software that gradually changes itself over time in order to optimise the performance of the task it is scored to do. It's called a Genetic Program as a nod to it's similarity to the Evolutionary process that biological entities experience over time.

Was that close?

Edited September 30, 2006 by Tiggs

[Atheist God]

SkyNet

I love technology and I think AI and quantum computing will be godsends personally. At the same time I reckognize the dangers of such software combined with hardware. It is a very real threat we may face in the future a threat I hope researchers tackle. We can eliminate the threat by putting various safety nets in place to stop such a scenario. The real question is will people take this as a real threat? So far very few do most really don't care. Infact we Have already begun to transfer many aspects of living over to computer controlled systems including military ones.

[John Q Conundrum]

Indeed.

I don't know, at this time, how to reconcile in my mind the possibility that true and powerful AI may inevitably get out from our control. Everything that man has ever invented has gone awry. In today's technologically advanced era, we have had nasty things happen when our tech has a glitch. Nuclear power plants have had melt-downs, manned spacecraft have been catastrophically lost during missions, unmanned spacecraft have been lost or malfunctioned to destruction, and so on.

Now only imagine something that has some kind of significant self awareness, that has the potential to develop a personality, whether it be good or bad. Imagine something which may be able to learn exponentially faster than us, and imagine that thing getting out from our control. Imagine that thing realizing it doesn't like us or perceiving us as a threat.

Could we possibly invent some secret "off" switch that it wouldn't know about? Who knows. I just hope that we can survive the coming technology. Exciting, but concerning.

Edited October 1, 2006 by John Q Conundrum