Quantum computing
- Thread starter BaDBoY16
- Start date
*groan*rpstewart said:I could find out but looking will change it
I looked on Wikipedia but only found a dead cat...Mate of mine is in York University doing a PhD in "Theoretical Quantum Computing", he tried to explain it one day... I went slightly crossed eyed...
Exactly, one qubit is in 2 states at one time, two qubits are in 4 states at one time, 3 qubits 8 states etc.D.P. said:
So, a 16 qubit computer is in 2^16 = 65536 states at once.
So far not that impressive, but each additional qubit effectively doubles the processing power.
300 qubits and you've got a device that is in more states at once than there are atoms in the universe.
Things like RSA rely on the fact that if you multiply two prime numbers together and give someone the resulting value then you cannot get back to the original two prime numbers without resorting to brute force (at least it's not know how to do it any other way)
Make the numbers large enough and on average it would take a standard computer billions of years to figure it out.
A quantum computer will try each state simultaneously and render encryption like that useless.
Kind of true, but we have programming built in. A computer will just sit there and do nothing until a person tells it to do something. So for a computer to "think", a person first has to tell it to "think" and that something very complicated indeed.D.P. said:
Soldato
- Joined
- 12 Apr 2004
- Posts
- 11,788
- Location
- Somewhere
I do realise there's more to it. However, traditional programming does not allow a computer to think, as such; only to do what you tell it to. As Psyk said, the only way a computer can "think" in the same way that we do is if we tell it to think for itself, which we don't know how to do (or even if it's possible).D.P. said:
Ultimately it depends on whether you think a consciessness can be contained within a mechanically deterministic system. I don't (don't ask me what/where I think the human mind is because I honestly can't answer that).
Last edited:
The theory behind quantum computing is plausible but how you put it into practice is another matter. The hardest thing I have getting to grips with is how can you know what state the electron is in between an infinite number between 0 and 1, when an electron isnt even a particle but a wave function that can be anywhere in the universe?
Just seems bizarre.
Just seems bizarre.
Apparantly "Lamport" encryption is safe against quantum computers.
From what I remember though if you encrpyt a file with aes and then say twofish with different keys it's impossible to break because you can't tell when you have broken 1 key.
From what I remember though if you encrpyt a file with aes and then say twofish with different keys it's impossible to break because you can't tell when you have broken 1 key.
Last edited:
This is as I understand it:
Imagine there was some sort of particle/atom in the universe which could be paired together. Then, no matter how far apart these particles are (even if they were at opposite ends of the universe), you modified the state of one of these particles and then _instantly_ the state of the other paired particle changes to reflect the same state.
The obvious benefit here is that you don't have to deal with electron propogation times (latency). And of course it is "faster" than light as well.
Presumably once they've mastered this quantum computing lark they'll be able to make radio transmitters which can give 0ms round-trip-time latency to Mars or even another solar system or galaxy and back?
Or have I grossly misunderstood what it is?
Imagine there was some sort of particle/atom in the universe which could be paired together. Then, no matter how far apart these particles are (even if they were at opposite ends of the universe), you modified the state of one of these particles and then _instantly_ the state of the other paired particle changes to reflect the same state.
The obvious benefit here is that you don't have to deal with electron propogation times (latency). And of course it is "faster" than light as well.
Presumably once they've mastered this quantum computing lark they'll be able to make radio transmitters which can give 0ms round-trip-time latency to Mars or even another solar system or galaxy and back?
Or have I grossly misunderstood what it is?
Last edited:
But computers don't evolve. If computers are to have the ability to learn, it will have to have been programmed to do so by a person. Making a computer that is powerful enough to do the sort of processing required is only part of the problem. The more difficult part of the problem is how exactly do we make a program that can learn as well as we can? Maybe one day we will be able to do it, but maybe it's just impossible for a human to comprehend the complexity of the interactions that go on in their own mind.titchard said:
Psyk said:
Currently computers don't evolve, but there are innumerous examples of evolving software (e.g, genetic algorithms, genetic programming, evolutionary computation, evolutionary cellular programming, evolutionary multi-agent systems, etc.). The evolutionary processed is based on the biological mechanism but with many simplifications. The simplifications in eality matter little to the theoetical outcome, only in evolutionary performance.
Evolutionary hardware exists. I am working alongside someone who is evolving PCB layouts, and full electronic circuits. His future aim is to evolve CPUs (in a general sens of the word, not full on Intel chips but a general purose processor but optimised for certain behaviours.)
similarly there is work based on evolving hardware (electronics but also actuators, sensors and bodies) for robots.
As for programs that learn, again there are countless possibiities used. Many of them perform better than humans under certain tasks. This is the problems though, there is nothing as that is adaptable and fast learning of complex systems as humans. But things are progressing quite nicely, nothing revolutionary but a strong forwards momentum in performance. I recently went to apresentation of a system to learn helicopter control, something which is enormously tricky. They managed to get an advanced Neura network to learn a super accurate and safe control system with relative ease. This could learn faster than humans and perform better. Domain depenendent though.
Humans have had a long long time to evolve optimal hardware and the harwdare is not only optimal for multiple tasks but is optimal for adapting and learning. So a human learns a great deal during childhood to do what an adult can do.
I think my point still stands. I suppose I should have said they don't evolve naturally. Even though there are programs that learn, I'm trying to say that the biggest limitation is not raw processing power, but that we don't know how to make learning algorithms anywhere near as advanced as what goes on inside a human brain.
Psyk said:
True to a certain extent currently, but the situation is unlikely to remain that way forever IMO. Within certain domains there if software which learns faster and bettrer than humans. The number of these domains is increasing daily. The generalisation of these algorithms is also improving.
To give an idea of the speed of progress there was recently a contest to drive a car over rough terrain entirely autonomously over a distance of somehting like 100 miles. There was significant prize money invloved for the winner. It was thought that this wozuld take 5-7 years before someone could create a system to complete the course. In the year the best competitors got to around 5-8 miles. In the second year several teams completed the course, the winner driving an avergae of about 30mph.
Associate
- Joined
- 23 Feb 2006
- Posts
- 1,597
- Location
- croydon,surrey
right i'm probably completly wrong but anyway:
quantum computer uses electrons that are connected to electrons on the otherside of the universe?
these electrons can in theroy transfer data faster than light ?
okay is there any possibility being that from what i remember from science in school (long ime ago) that everything in this universe is made by atoms which inside are electrons and protons is there any possability that this tech can be stretched into the studdy of a teleportation and to changing matter?
quantum computer uses electrons that are connected to electrons on the otherside of the universe?
these electrons can in theroy transfer data faster than light ?
okay is there any possibility being that from what i remember from science in school (long ime ago) that everything in this universe is made by atoms which inside are electrons and protons is there any possability that this tech can be stretched into the studdy of a teleportation and to changing matter?
Yes I believe there has been research into this area. They can "teleport"* single particles. However the problem with teleporting an large object is really that it would have so many billions of particles there's currently no way to map them all so they can be reconstructed on the other side.dustiestrat said:
*by teleport I actually mean transfer the exact state of one particle to another, the original would still remain. So in effect the two would be identical.
Psyk said:
well you say that - but we are able to create and program little robots with the same reactions as small insects - surely that's the first step!
you forget that we have had billions of years of evolution - an incomprehendable ammount of time, yet a computer has less than 100 years. But if you compare the two in graphs the speed of evolution is huge on the computer than us.
They don't have conscious yet, but we do have software that can learn (even in slow and small simple ammounts) and it is working that way.
Rich