how is data encoded in quantum computers? I know that qubits store data as a probability to be 1 or 0 but and you can use that to store multiple values in the same space but how exactly is that encoded
for example, would a list of every single possible byte be stored as just a qubyte where every qubit has a 50/50 chance of being 1 or 0?
>>54539004
bumperino for interest
why is there no widespread information about quantum computing
Seems all secretive
>>54539121
this
i just know about how a quantum computer is supposed to work but i have no idea how the implementation could work
>>54539004
There are many different approaches to quantum computing. First thing to remember that it is probabilistic. You perform a calculation and you will have a certain probability of measuring a correct result. There are many different approaches to quantum computing. One tangible example could be using the spin of an electron. You could prepare the spin of an electron in a up ( the 0) or a down state ( the 1). You could also store it in a superposition of 0 and 1. This electron can store two bytes. With n qubits, you can encode 2^n bytes.
Basically a qubit is a quantum particle that encodes some quantum information. Due to the wavefunction of certain things, like electrons and other subatomic particles, they are in what we call a superposition state of two or more states.
There is a way of generating a pair of electrons with different spin. One MUST be spin up, the other MUST be spin down. If the probablilities are uniform, one of them is in a superposition of 0.5up | 0.5down.
Now imagine if we have some sort of quantum gate that encodes a number from 0 to 1 and puts that value as the probability that an electron is spin up and puts out that electron.
That's what they mean by a qubit. Each bit can store a ton of information.
The problem with quantum computing is that right now if you want to observe the output of a qubit, you have to destroy it. The problem is you can't observe the probability; you can only observe samples of the distribution. So if you wanted to check whether or not the probability is >0.5, you would take a whole bunch of those electrons generated from the gate and check how many are up and how many are down. The great thing is the longer you run the computation the more accurate your output gets. (assuming your primitives like gates are perfect)
RAP BATTLE BETWEEN RMS AND DRAKE AT >>54538692
wew
>>54539190
kill yourself
http://www.rintonpress.com/journals/qic/7-8-799-800.pdf
It's "secretive" because the internet sucka as a medium. The only good thing about the web is the exchange of information, but the actual information is mostly in books. That's a one page review of a nice book on quantum computing that serves as an intro, you'll find your answers there. Read it and if you're interested get the book
>>54539175
Literally everyone knows that part. I'm asking how you would encode a set of possible values in qubits
say you have a set of bytes [1, 52, 12, 147, 89, 202]
how would you store all those different values in one qubyte? What are the states of the individual qubits?
>>54539277
You would still use a register of qubits, except each qubit can store 2 bits.
>>54539265
What about jargon? I want to understand this even if I don't necessarily know the terminology. I've struggled learning quantum physics because while wikipedia does provide answers, the answers are so technical I that they're functionally useless
>>54539302
Don't even try to understand quantum computing without the equivalent of an intro course to quantum physics and a good knowledge of linear algebra.
Why the fuck are you learning from Wikipedia when open courseware exists?
>>54539277
You have to think differently. A quantum computer is not a replacement for a classical computer. You could store specific byte strings in qubits, but you just don't do it. The neat thing about quantum computers is that you can have every possible state at once. For example you want to find all primes between 0 and 255. You just take 8 qubits, which are in all those states simultaneously. Then you apply a bunch of quantum gates on those gates that will eventually lower the amplitudes of the non-prime states until they are practically zero. Then you look what kind of state you get.
You can basically get rid of all iterations in classical computers. You don't need to test every single combination of bytes successively, you can actually do them all at once (usually you need to do that multiple times to get all answers though).
>>54539277
>say you have a set of bytes [1, 52, 12, 147, 89, 202]
>how would you store all those different values in one qubyte?
putting aside this rather rough treatment of the term "byte"
if we stuck with a 0-255 "bit" range, assuming equipment capable of measuring quantum distribution at that resolution, encode each element of that array in a qubit.
[1 52 12 147 89 202] becomes (example) spin state distributions of
[0.39% 20.3% 4.7% 57.6% 34.9% 79.2%]
where in each case you'd interpret the information as "x% likely to observe a spin-down state over a spin-up state." Contemporary tech would have you put at least a few hundred/thousand runs through any logic in order to estimate the distribution by sample probabilities. Future quantum tech would have us observe those probabilities directly without sampling.
If we have greater resolution, say 0-65535, we could bytepack them into a smaller number of qubits instead. But I think that's not really the point of your question?
>>54539297
if it were that simple, quantum computing would be worthless. Isn't the point storing huge amounts of data in different co-existing states of the same bits and operating on them all at one?
>>54539331
I took linear algebra 1 before I dropped out of the physics program at my school and I've ben reading up on the basics on my own. I still don't know most of the formulae but at the very least I understand how superpositions work. They're just complex vectors for which |a|^2 + |b|^2 + ... + |n|^2 = 1
>>54539344
I get that part but I don't really get how all those separate states are encoded when the bits that make up the data are independent
Like, if you had two qubits and they stored a superposition of 10 and 01, how would you collapse them without worrying that they could come out 00 or 11?
>>54539410
>I get that part but I don't really get how all those separate states are encoded when the bits that make up the data are independent
>Like, if you had two qubits and they stored a superposition of 10 and 01, how would you collapse them without worrying that they could come out 00 or 11?
That is literally the magic of quantum mechanics. Mathematically you store the state of the n qubits in density matrices. If you want to understand that in more detail, you NEED to take a course in quantum mechanics and even then it's a lot to buy.
In the end it always boils down to "that's just how it is m8".
>>54539431
Oh, so instead of thinking of it as a vector getting longer, I should think of it widening into a matrix where the chance of one qubit being in one state depends on the state of every other qubit?
>mfw I don't know dick about quantum computers other than they will utterly fuck up computer security
evidently quantum computers make damn near everything about computer security worthless
>>54539553
Is this a question?
>>54539553
maybe you could have some sort of hybrid system with separate memory, storage and processors and design it in such a way that passwords can only be accessed through a classical computer interface.
Important information data going from the quantum half to the classical half would have to be collapsed and data going from the classical to the quantum would be stored in such a way that they aren't superpositions.
This would force password and key stuff to happen at classical rates while actual computation could run and quantum computer speeds
>>54539553
There's existing encryption algos that are not susceptible to quantum computer attack, though.
There's very few algorithms where quantum computers beat classical ones in complexity. Integer factorization is one of them.
>>54539121
>>54539146
Because there aren't widespread quantum computers yet available. There isn't a single common implementation, apart from D-wave all QC are in labs and use only a handful of qbits.
The qubits can be implemented in various ways, pretty much any harmonic oscillator can be used as one, NV-centers in diamond, trapped ions, quantum dots, superconducting ring currents... And each of those things can be implemented differently, depending on the requirements - room temp/helium cooled, milli-/microseconds of storage lifetime, noise tolerance, precision, cost...
This is literally everywhere if you know what you want to know, wiki for the broad picture, arXive for state-of-art details
If you are happy to leave the messy details to engineers, you "only" need to be comfortable with Linear algebra. Each qbit is a complex vector (a,b) with a*a+b*b=1. The experimental setup is a hermitian matrix and a projective measurement yields one of its eigenvalues with a probability equal to the scalar product between the qubit and the corresponding eigenvector. That's it, go ahead and think of countless ways to store information in one or more qbits. Now add 15%noise, boundary conditions and entanglement and do it again
what will be the first game to be made for a quantum computer?
>>54539731
Quantum computers would be much shittier than a current laptop for games.
>>54539788
could better AI for games be made/run on it?
>>54539731
Pokemon Space and Pokemon Time!!
>>54539801
certain applications would be better. For example, a quantum computer would be much better at pathfinding. Say you can only move one space in any of the four cardinal directions. to figure out how to get to a point on the grid, a classical computer would have to try a number of sequences equal to 4^n where n is the smallest number of steps needed to get to that space. By contrast, a computer would only need to test n sequences.
For the record, those two are incomparable
>>54539004
>>54539121
https://youtu.be/ZoT82NDpcvQ
I got this out of a google search
OP, have you tried searching?
>>54539004
This video may help you understand the basics.
https://www.youtube.com/watch?v=1PcseLsYZ9Y
It's very, very very different. It's probabilistic. There are things you cannot do with a quantum computer. It is not known if there is an upper limit of the number of particles that can be entangled with each other.
And it will probably always need heavy supercooling - not for superconduction reasons, but because otherwise the thermal noise drowns out the measurements.
>>54539962
So lazy evaluation would be based in a quantum computer
A bunch of particles are in "all" states.
Then after several operations; these states are reduced/changed to a different set- invalid states collapse and disappear from the probability.
Following this, you can design a sequence of operations / program that should return the only valid state left.
You can then observe the particle (it'll give the state you want, probably) and can then reverse engineer the algorithm to find which input gave you the answer you wanted.
That is to say, solve a problem backwards by simultaneously bruteforcing all possible inputs.
>>54539553
This is false. It only completely breaks factorization and discrete log.
Symmetric ciphers are safe, lattice algos are safe, etc.
Post-quantum cryptography is a thing, and we should all be researching it right the fuck now because the data we store TODAY needs to be resilient against the methods of the FUTURE.
>>54541369
You sound like a post-cold war commercial.
>>54539676
What stops someone copying the classical memory to their quantum computer and solving?
>>54541369
Almost. Symmetric ciphers are safe if we double the keysize (256 is fine); hashes if we triple the output size (384/512 is fine; 256 is too short). Don't forget Shor's algorithm.
It's currently unknown whether ideal lattice-based cryptography is safe even on conventional computers. (Of course, this is also true of RSA. We don't know that a fast factoring algorithm doesn't exist in O(n).)
If you've seen Campbell-Groves-Shepherd (2014) "SOLILOQUY: a cautionary tale", and Albrecht/Bai/Ducas you'll also know that the structure of the rings (usually cyclotomic) used in Ring-LWE and NTRU may allow for practical attacks - again, not requiring a quantum computer.
We need to take great care that, in defending cryptosystems against quantum computers, we don't make something worse than the cryptosystems we already have.
djb/Tanja Lange and two others have recently proposed NTRU Prime, which uses a prime-degree, large-Galois-group, inert-modulus ring. It's also handily faster than Curve25519. (Although the keys are bigger - 1232-byte public keys and 1141 byte ciphertexts, compared to 32.) https://eprint.iacr.org/2016/461
>>54541707
>Shor's algorithm
Sorry - I meant Grover's. Shor's is integer factorisation. Grover's can speed up any brute-force search by a square factor: 2^n -> 2^(n/2)
>>54541707
>It's currently unknown whether ideal lattice-based cryptography is safe even on conventional computers. (Of course, this is also true of RSA. We don't know that a fast factoring algorithm doesn't exist in O(n).)
Aren't there multiple lattice cryptosystems that are provably NP-hard?
>>54539004
you can now use IBM's 5 qubit quantum computer via the internet
http://www.ibm.com/blogs/think/2016/05/03/the-quantum-age-of-computing-is-here/
>>54541843
>we created a web app running on an x86 processor that simulates a quantum processor
wew lad
you should try and get a bunch of random numbers out of it and graph the patterns, maybe you can even find a telltale sign of a PRNG :^)
>>54541916
Dude you can actually run it on their quantum computer
http://www.research.ibm.com/quantum/
Can we emulate quantum computers with classical computers?
>>54541963
Of course. It just has exponential runtime.
