quantum computers?

[reyam200]

ive always thought it would be cool to have a computer that thought faster than any human, and lately computers are continuing to get faster and smarter. When do you think quantum computers will be made? ones that your average person could buy.

 

id love to have V.I.K.I from the movie "IRobot" although not have here connectid to internet all the time. just as a precaution.

 

 

also with the existence of quantum computers, it would be possible for really advanced AI. since QCs can preform many tasks at once.

[5614]

You like computers which can think faster than you? And you are looking at quantum computers? You do realise that the average home computer can do many more millions (or billions) of calculation per second compared to you, does that not count as thinking faster?!

 

As for quantum computers... well the prototypes are there and working, but it is a very long way off commerical use, even further from home use. If you consider the first computer to be made around about World War 2, only in the late 90s did we start seeing home computers becoming common, that's about 50 years. If quantum computers were properly starting to be developed a few, maybe 10 years ago, then you could argue by 40years they will be ready.

 

Of course you can argue that time to make a conventional computer and time to make a quantum computer are totaly unrelated, and indeed I just made that up. The point is that no one really knows the answer.

 

There are working prototypes but there has been no major sudden breakthrough allowing it to be used at home, it is very complex and requires many many years of hard work.

 

Check out IBM's Physics section: http://www.research.ibm.com/disciplines/physics.shtml

[Nacelunk]

We can create computers that perform calculations trillions of billions times faster than we do but we will never ve able to create intellegent computer until we fully understand how our brain works. And I think that would be much more important event than creating super fast computer

[JustStuit]

There is an article in this month's Scientific American about computing with quantum knots. One trouble seems to be keeping them isolated because errors can occur when not properly stored.

[bombus]

We can create computers that perform calculations trillions of billions times faster than we do but we will never ve able to create intellegent[/b'] computer until we fully understand how our brain works. And I think that would be much more important event than creating super fast computer

 

Oh can we now! I think that rather depends on what you consider to be a calculation. Robots, even those connected to supercomputers, have major difficulties performing what we consider to be simple tasks such as walking, or bending down to pick up an object without falling over let alone anything truly athletic. Surely the calculations brains perform to enable us to do these tasks would outstrip any computer? I think there is a football match planned for 2050 pitching robots against humans. I think the humans will win!

[RyanJ]

I agree with bombus.

 

Computers may calculate faster but they have less power, if they ever develop bio-computers (which I am shure will be developed in the next 20 years) these will act as the brain does in array sets rather then direct linear links.

 

Even those these are slower their power is much greater, as with the brain its slower but much, much more powerful than any computer on the planet.

 

Computers also lack the ability to adapt, or alteast they can do so in a very limited manor, this would be more easily developed (an adaptive framework) with with bio computers.

 

Computers are currently better for one thing right now, speed. There is no computer that can understand yet so its going to be an interesting time these next few years considering the considerable progress made in the last decade but we are still far away from a V.I.K.I style system yet

 

Cheers,

 

Ryan Jones

[5614]

Obviously a computer is amazing at calculations and terrible at doing basic, almost unintelligent things which a baby could do, such as recognising a person (a baby would recognise their parents).

 

Also an issue human-computer interaction is very different from human-human interaction in a very deep and subtle way. For example (I had a lecture on this the other day) the lecturer asked a member of the audience if he had the time, the guy replied "yes, it is 10 past 2". The lecturer said "exactly the point... if you asked a computer if it had the time the computer would say 'yes'".

 

The lecturer never asked what the time was, but it was obvious he wanted to know that, that is the difference between human-human and human-computer. Another example would be if you were hungry you could ask a friend to go get a chocolate bar, and they would go to the shop and buy a chocolate bar and bring it back for you, and all that you did was say "I want a chocholate bar". But for a computer you would have to say 'walk forward 5m, turn left, walk forward 200m, pick up chocolate bar located infront of you, pay, turn around and a whole journey home. When talking to a computer a simple "get me a ___" statement becomes very complex. Making a computer understand what a statement (e.g. "get me a ____") really means (e.g. walk to shop, buy, return and how to do those etc.) and get it to do that.

[Sexington]

But Computers can do amazing things that humans would have no chance of doing without them. It seems like nowadays we depend on these machines to function. I don't think we'll see the day that we can develoup an AI to act, think, and actually become like a human being but without them, many things wouldn't be possible.

[5614]

Computers can do amazing calculations that humans could never do, agreed.

 

What we are pointing out is that there are also things which a baby could do and a computer couldn't, such as recognising a human face.

 

Although maybe that is not so fair. The lecture I had the other day on AI and multimedia searches showed an example where the lecturer placed a picture of a blue door into the search box, using the pixels alone (ie. not the image name) the computer could search a database of thousands/millions of images and return other images which the computer thought were related. This mainly included doors, some of which were blue. Upon specifying the search further (by selecting multiple blue door pictures) the computer realised that what we were searching for was a blue door so returned again many doors, this time the majority of which were blue.

 

Similarly the computer could differentiate (using the pixels of the image only); sunsets, grass, trees, buildings, water, sky, if it was an areial image or not and that was just an example or two he showed us, I'm sure the computer could identify other features.

 

If this kind of research went further the idea is that one day a computer could recognise a face, then expressions etc. etc.

[ashennell]

Quote:

Originally Posted by Nacelunk

We can create computers that perform calculations trillions of billions times faster than we do but we will never ve able to create intellegent computer until we fully understand how our brain works. And I think that would be much more important event than creating super fast computer

 

Oh can we now! I think that rather depends on what you consider to be a calculation. Robots' date=' even those connected to supercomputers, have major difficulties performing what we consider to be simple tasks such as walking, or bending down to pick up an object without falling over let alone anything truly athletic. Surely the calculations brains perform to enable us to do these tasks would outstrip any computer? I think there is a football match planned for 2050 pitching robots against humans. I think the humans will win![/quote']

 

Genius. Did you bother to read nacelunks point!! You agree with him. His point - even if we could make really fast computers they won't be intelligent unless we can understand the types of calculation made by the brain.

[ashennell]

If this kind of research went further the idea is that one day a computer could recognise a face, then expressions etc. etc.

 

Unsuprisingly, this is a already a major research area. Although I think that results so far have been disappointing.

 

http://www.face-rec.org/ - a good place to find out more.

[RyanJ]

The idea is a general one, recognition is simple comparative recognition is a pain.

 

Computers have to be told exaly what to look for so with applications where there are variations then it is much harder.

 

There are much more complex things and I agree we will not be able to replicate them untill we understand then - I'm not shure we will even then. Emotions are one great example - we are not shure if its a specific process or something related to everything else.

 

Like I said it will be an interesting next 20 years

 

Cheers,

 

Ryan Jones

[Cloud]

With quantum computing(with data on atomic scale) - wouldn't you be able to just hook up to a human brain and then (with the power of a scientists brain combined with the calculative power of a computer) be able to fully understand how the brain works (as someone said earlier).

[5614]

Err, basically no. We could not hook a computer to a brain for the same reason we can't do brain transplants. We don't know enough about the brain. We wouldn't know where to connect it to, even if we did connect it what would we get? A bunch of electrical impulses.

 

As for what RyanJ was talking about, sensing an object is easy, comparing it or understanding what it is is harder. My post #9 is very much related to that. So ok, now a computer can determine what it in an image.

 

Another thing we were shown was a text based search for images, using only pixels. So the guy searched for "scenic panoramic views" or some such and it came up with mountain ranges etc. the point is that it was taught how to recognise features of an image, it was then fed thousands of images which it automatically recognised and could then return relevant results when we search for something specific using only the pixels in the image. A lot of progress is being made, but it is a long way. Recognising the image and interpreting humans (what I talked about in post #7) is a big step.

[RyanJ]

Err, basically no. We could not hook a computer to a brain for the same reason we can't do brain transplants. We don't know enough about the brain. We wouldn't know where to connect it to, even if we did connect it what would we get? A bunch of electrical impulses.

 

Yea that would be like trying to hook up a modem to an ariel and expecting to see a TV picture on the computer.

 

There would need to be something to turn the neuron's pluses into somethingthe computer can understand - just like the modem does for the "noise" over a phone line Rightnow we have no idea how we could possible do that

 

Cheers,

 

Ryan Jones

[dcstegg228]

Brains are far superior in the functions thay can perform at the same time without slowing down. and computers cant hold anywhere near the amout of a human brain.

[5614]

I beg to differ. A computer I saw (ie. in real life, not a picture) just yesterday happened to have 90TB (that is 90,000GB!) of storeage space, that sounds like a lot more than I could hold!

 

NB: The computer also had 400 individual 64-bit processors, 400GB of RAM etc. etc. btw, it wasn't mine, shame, but there ya go!

 

Also my computer which can handle several million calculations per second is a tad quicker than my brain, maybe not yours, I wouldn't know!

[dcstegg228]

I beg to differ. A computer I saw (ie. in real life' date=' not a picture) just yesterday happened to have 90TB (that is 90,000GB!) of storeage space, that sounds like a lot more than I could hold!

 

NB: The computer also had 400 individual 64-bit processors, 400GB of RAM etc. etc. btw, it wasn't mine, shame, but there ya go!

 

Also my computer which can handle several million calculations per second is a tad quicker than my brain, maybe not yours, I wouldn't know![/quote']

 

Maybe calculations in math but your brain uses only 5% of its power to controll evbreything you do at any given time. and i think the brain holds more than 90tb

 

maybe it holds like 1pb or sumthin (1024tb) and 90tb is 92160gb

[5614]

Yeah, technicalities because 1KB = 1024B but it is always simplified to 1000, yeah, we all know.

 

I don't believe a brain could hold 90TB of information, but I guess it's not really something you can prove or disprove and I don't really want to get into a debate about it.

[Cloud]

Someone of my level of science shouldn't bother but here goes ~

 

Wouldn't we be able to detect (using Quantum computers) different emotions produced from the brain through these electrical impulses (changes at atomic level) and then upload the correspoding qubits onto a QC as a program and thus given the attributes of emotion to a computer.

 

Please clarify why the above is ridiculous.

Thanks

[RyanJ]

I don't believe a brain could hold 90TB of information' date=' but I guess it's not really something you can prove or disprove and I don't really want to get into a debate about it.[/quote']

 

We also would need to know wha a kilobyte is in terms of organic memory for example, we are talking about two completly different systems so how can we even compare measurements?

 

Savants are an example, one person, Kim Peek can remember everything he has ever read. He can recite these books work for word and he reads afew new books every week not to mention huge amounts of dates and figures too.

 

That type of memory is amazing and everyone should be able to do it because his brain structure is almost idential to that of your average person.

 

Who knows!

 

Cheers,

 

Ryan Jones

[5614]

Savants are an example, one person, Kim Peek can remember everything he has ever read. He can recite these books work for word and he reads afew new books every week not to mention huge amounts of dates and figures too.

Do you realise how much text you could store on 90TB? I guarantee it is more than you could possibly read in an entire life time.

 

Wouldn't we be able to detect (using Quantum computers) different emotions produced from the brain through these electrical impulses (changes at atomic level) and then upload the correspoding qubits onto a QC as a program and thus given the attributes of emotion to a computer.

No. If we plugged a brain into a computer, not that we know how to, we would just get a bunch of electrical impulses (that is how the brain works). We would not be able to interpret what these seemingly random electrical impulses mean. The electrical signal would be more like a bit than a qubit. A QC may work differently inside to a normal computer but fundementally it does the same job (just using a different technique to do it), there is no reason why this should have to be on a QC as opposed to a normal computer.

[RyanJ]

Do you realise how much text you could store on 90TB? I guarantee it is more than you could possibly read in an entire life time.

 

Yes I know how much it is... its not only text its other things, audio, picture and even moving pictures.

 

Also consider that the brain physially makes connections while a computr simply puts a 0 or a 1 for each bit on a disk or in a RAM slot.

 

I'm not saying the estimate presented here was close but I am saying is a lot more then we probably believe it is

 

When you think of it in the prespective of the shere volume of data stored it makes my headhurt, all that information accumulated over an entire lifetime... its a lot.

 

Cheers,

 

Ryan Jones

[ashennell]

Also my computer which can handle several million calculations per second is a tad quicker than my brain, maybe not yours, I wouldn't know!

 

Most peoples brains can do alot more than that!! Each neuron may be slow, maybe no more than 100 calculations a second but, we have a lots. I think somewhere around 100 billion.

 

I don't believe a brain could hold 90TB of information, but I guess it's not really something you can prove or disprove and I don't really want to get into a debate about it.

 

I agree - It's pretty much impossible to compare. One way that is often used is to estimate the number of synaptic connections and use that as an estimate of the number of free variables.

 

Maybe calculations in math but your brain uses only 5% of its power to controll evbreything you do at any given time. and i think the brain holds more than 90tb

 

5% of it's power? I'm not sure what this means - it can do things 20x faster but never bothers? It this just a variant of the old 'we only use 10% of our brains' myth?

 

 

Wouldn't we be able to detect (using Quantum computers) different emotions produced from the brain through these electrical impulses (changes at atomic level) and then upload the correspoding qubits onto a QC as a program and thus given the attributes of emotion to a computer.

 

I don't know anything about QC's or qubits but I do know about brains. We don't need to look any deeper than neurons and neurotransitters to understand emotions. Emotions will eventually have a computational explanation, possibily related to grounding of internal models or enforcing some variable constraints on the heuristic processes that brain uses.

 

Transfering data from the brain to the computer to make it have emotions is implausible. Any recorded information only makes sense in the context it was recorded - if you want to to take on the same semantics. I'm not really sure what a qubit is to I can't say more.

[Klaynos]

A mildly interesting comparisan to make is that of neural networks in computers. And how even a simple one mapping just a few neurons takes large amounts of computer power due to the way they change and form links etc...