A "Photon Machine Gun" For Quantum Computers

An anonymous reader writes "Generating entangled photons in a reliable way is impossible right now, stalling the development of the optical quantum computers that would use entangled photons as quantum bits (qubits). Because entangled photons can only be produced at random — which takes time — the most powerful optical quantum computing device use only 6 qubits. UK and Israeli quantum physicists have designed a blueprint for a 'quantum machine gun' that fires out barrages of entangled photons on demand. They think within a few years this device will be built, and could lead to quantum computing using 20 to 30 qubits. Every additional qubit doubles the computing power, so these quantum computers could outperform any existing classical computer, the researchers say. The quantum machine gun is described as 'one of the most exciting theoretical proposals I've read in five years' by a leading quantum physicist." The research was published in Physical Review Letters earlier this month.

Comments

For certain problems.

[JorDan Clock]

Every additional qubit doubles the computing power, so these quantum computers could outperform any existing classical computer, the researchers say.

But only for probabilistic algorithms. It's not going to be faster at everything.

Re:For certain problems.

[Tablizer]

But only for probabilistic algorithms. It's not going to be faster at everything.

So Whpt if we occjsion?lly fl#p a fwe bits.
   

Re:For certain problems.

[maxwell demon]

As far as a know it not yet known if a quantum computer can turn NP complete problems, in polynomial problems at all, or for what problems this is possible.

Of course, as of yet it isn't even known if a classical computer can calculate NP complete problems in polynomial time. P!=NP is still a conjecture.

BTW, the correct arXiv reference is arXiv:quant-ph/0601151. After all, there's also astro-ph/0601151, cond-mat/0601151, hep-ph/0601151, hep-th/0601151, math/0601151 and physics/0601151, none of which are relevant here.

Qubit does not double power in traditional sense?

[religious freak]

Every additional qubit doubles the computing power, so these quantum computers could outperform any existing classical computer, the researchers say.

I thought that the "power doubling" was not in a traditional sense.. the qubit is fantastic at pattern matching and search functions, but no better than a classical computer for something like, say, a video game requiring finite mathematical calculations. I'd state this as a fact, because I've read this in at least a couple places, but seeing as how quantum physicists haunt this forum, I can't say I know as well as them. But this power is only useful in very specific circumstances, AFAIK.

Not particularly useful against an insurgency

[BadAnalogyGuy]

Israeli quantum physicists have designed a blueprint for a 'quantum machine gun'

In other news, Palestinian quantum physicists have designed shoulder-mounted quantum launchers and quantum vests in response.

Civilians are hopeful for peace and terrified for escalation of hostilities.

Re:Not particularly useful against an insurgency

[Shadow of Eternity]

Harmful harmful force? Dude i think you need to re-evaluate your worldview if you want to blame the group being constantly attacked and threatened with the explicit goal of genocide for everything wrong. The mere presence of jews in the middle east produces the reaction you see from Hamas and friends, whether or not Israel was officially a state would have fuck all to do with anything other than the success of those attempts at genocide.

Hell Hamas' own govt charter explicitly blames the jews (merchants of death) for everything from the french and russian revolutions to both world wars while outright demanding the death of every jew and anyone who refuses to participate in said genocide.

Re:Not particularly useful against an insurgency

[Ma8thew]

If Israel wishes not to be treated as the bad guy, then they should stop acting like the bad guy. Demolishing people's homes, closing borders to make Gaza a prison and handing down collective punishment do nothing for Israel's image, and gives the religious extremists in the region a concrete reason for the destruction of Israel. If you think Israel are the innocent victims of Palestinian oppression, then you really need to re-evaulate your worldview.

Re:Not particularly useful against an insurgency

[Shadow of Eternity]

So in other words you want them to do the exact opposite of the only thing keeping the number of rockets and mortars fired at Israeli (arab AND jew) civilian targets in the low to mid thousands rather than much higher.

What did we do with vietcong tunnels in vietnam? The ones used for moving and storing weapons and occasionally fired from. We demolished them. What does israel do with houses build on top of tunnels or used as weapons caches. They demolish them. If people don't want their house bulldozed all they need to do is say "No you cannot indiscriminately attack civilians from my house or store/transport the weapons you use to do that in/through my house."

As for the borders, wtf do you suggest they do? Just open them up for MORE weapons to get smuggled through?

The religious extremists in the region already HAVE a concrete and unarguable reason for doing what they do, it's called genocide. This has NOTHING to do with Israel as a state and everything to do with the fact that there are jews and christians (but mostly jews) over there that aren't dead yet. Read the Hamas charter sometime, the slaughter of all jews everywhere is listed as mandatory for the messiah to come in it. If you think anything Israel does has ANY bearing on anything the palestinians do you're delusional.

There's a palestinian couple living a few apartments away from me, do you know what he calls palestinians that don't want to kill all the jews? Israelis. Just like the million and a half arabs living in israel that are ALSO a target of palestinian violence because they don't join the genocidal crusade Hamas is currently leading.

Re:Not particularly useful against an insurgency

[cheesybagel]

If you go further back in history the Jews are supposed to originate from ancient Sumeria. This is one of the reasons they incorporated the Epic of Gilgamesh into their own mythos. With all historic movements nearly no one is from the place they are at.

Re:Not particularly useful against an insurgency

[Shadow of Eternity]

http://lmgtfy.com/?q=hamas+charter

Re:Qubit does not double power in traditional sens

[TubeSteak]

I thought that the "power doubling" was not in a traditional sense.. the qubit is fantastic at pattern matching and search functions

Which is all that really matters for breaking encryption, and is the whole reason we have computers in the first place.
So my question is how many bits of encryption do I need to keep a 20~30 qbit computer out of my truecrypt partition?

no peeking

[RuBLed]

It's not going to be faster at everything.

It's going to be faster at everything.

Re:no peeking

[Anonymous Coward]

It's not going to be faster at everything.

It's going to be faster at everything.

It's going to be simultaneously faster and not faster at everything.

Re:no peeking

[plastbox]

One (*err* more) thing I don't get.. How do they know quantum entanglement even happens? They entangle a pair of particles. Then they measure the state of one, causing the other to collapse into the same state with no regard to distance between the two.

However, as it is impossible to measure the quantum properties of these particles without collapsing them into a non-super state, how do we know that the entanglement wasn't just the two particles gaining the same properties at the moment of entanglement? Obviously, this would result in them having the same properties once measured.

How do we know that this super state exists, when it is impossible to measure? If a piece of equipment paints two balls a random color and puts them in separate boxes aren't the balls, by the same definition, in a super state as we can't know their color until we open the box? And can they be said to be entangled, as once you open the first box and observe that the ball inside is for example red, the other ball will also be red even though it has yet to be "measured"?

This might be a bit of a Captain Obvious statement, but I don't freaggin get it! =(

Re:no peeking

[jellomizer]

Well... it is both.

Re:no peeking

[Artifakt]

First, let's look at a fair attempt to explain why quantum indeterminacy is not just the same thing as classical indeterminacy (like your two particles, which by your question were presumably determinate in the classical model, at least until they became entangled). You seem to be reasoning much as the following note claims early quantum physicists tried to, when they first grappled with Heisenberg's uncertainty principle and the question of knowing the position and velocity of an electron simultaneously. I give you someone deliberately trying to put the concept in normal, natural language and not use any actual math:

http://www.uhh.hawaii.edu/~ronald/310/Quanta.htm

One point is, the interpretation that we can't know both position and velocity at the same instant, therefore the electron doesn't have both at the same instant, doesn't explain that thing you refer to as "with no regard to distance". This is what sometimes gets called "Spooky action" and is related to non-locality in general. Starting from the interpretation that it's not our not knowing that causes the indeterminacy but the indeterminacy which causes our not knowing turns out to be putting the horse back in front of the cart. Once people started working from the idea that the indeterminacy is fundamental and not like your example of the balls (where there is a definite color for each, and the observer just doesn't know it yet), they started making progress on figuring out how entanglement could be faster than light.

http://www.absoluteastronomy.com/topics/Quantum_indeterminacy

This is about what non-locality really means: One consequence is that we can't assign a local cause (such as: a localized observer hasn't looked yet) to explain why something on the quantum level is determinate, or we lose the ability to explain how the faster than light part happens.

Just as the original QM problem was about determining position and velocity, talking about "non-localizable" (position), and instantanious/faster than light (velocity) is two ends of the same stick. The more you prove that the action happens much faster than the limitation of light-speed, the more you can't claim the action is caused by anything in a particular locale.

Re:no peeking

[wurp]

OK, you measure photon phase with a polarized lens. The way you measure the phase is to pass the photon through a polarized lens at an arbitrary angle. Unfortunately, all you can measure about the phase is whether the phase matches that of the lens - either the photon makes it through the lens, which means the photon had the phase of the lens, or it doesn't, which means the photon had a phase at 90 degrees to the lens. There's lots more to say about this, but I think this is enough to explain the answer to your question.

If you measure the phase of a photon with random phase (whether it's just classical random meaning it has a phase, but you don't know what it is, or if it's quantum random, meaning it's unknowable), half the time the measurement will say it's in phase, and half the time it will say it's 90 degrees out of phase. It will never say anything else (obviously).

If you have two photons with entangled phases, and you set up lenses at 90 degrees to one another, you will always (guaranteed, at least up to how accurately you can get the lenses at true 90 degrees to one another) get corresponding phase readings. Note that the actual angle of the two lenses doesn't matter - just their relative angle.

If the entangled photons had an initial phase, there's no way that could happen. If the angle happened to be 45 degrees off from your lenses, each of the photons would only get through half the time, and both would be random, not correlated. Either the photons somehow "know" what angle the lens you're going to use to measure them is at the time the photons are emitted (even though the lens might not be set up then), or they "communicate" about what angle the lens is after they hit it. (Or they have an infinite number of correlated states telling them how to react to every possible angle, or...)

In any case, the phase is not the expected classical relation.

Re:no peeking

[CopaceticOpus]

The relevant theory here is Bell's Theorem (or Bell's Inequality.) The principle of entanglement has been shown experimentally using some clever approaches based on probability.

If you measure a specific property of two entangled particles, you are correct in saying that there is no way we could know if the result of the measurement was predetermined. However, experiments were set up in which a large number of pairs of particles were measured. Each measurement recorded one of several possible properties, chosen at random. Sometimes the same property would be measured for both particles, and sometimes different properties would be measured for each particle.

It can be shown mathematically that if the particle properties were predetermined, we would see certain probabilities emerge. In other words, for each pair of measurements, the same result would be found N% of the time. However, if the particle properties are determined at the time of measurement, the math changes, and we expect instead for the results to correlate P% of the time. This latter result is what was observed.

That is just a quick overview of the concept. I suggest reading Brian Greene's "The Fabric of the Cosmos", which provides a great explanation involving Mulder and Scully. (Really!)

I also found a nice, clear explanation of the probabilities involved here.

ObCompJoke...

[Tastecicles]

imagine a Beowulf cluster of... NO! NONONO!

Re:ObCompJoke...

[Yoozer]

I tried to imagine a cluster of photon machine guns and all I could come up with was a container full of Mag-Lites.

Explain the hype, please?

[plastbox]

Ok, so on this site bursting with intelligent, educated folk, the following question(s) might make me look like a village idiot, but what the hell. It's damn interesting stuff and I want to know!

Exactly how does quantum computing work? I have a fleeting grip the basic stuff; qubits existing with states 0, 1 and "superposition" (i.e. all possible states) and that by actively measuring it's state (sending a photon or whatever bumping into it) you collapse it, and it's entangled mate, into a "classical state". If I place a shot glass in a dark room and tell you it could be empty, full or anything in between but the only way for you to find out is to A) Take the shot, or B) dump another 4cc of Tequila into the glass and see if it spills over, is the shot glass a cubit? To you, it is in a "superstate" until you actively measure it, an act that in itself makes the glass full or empty.

How does this equate to computing? I might just have spent too much time with Proteus fiddling about with gates and stuff trying to make a very basic functional computing device, but isn't some sort of computing device needed to compute something? Even with Quantum Gates, 30 qubits seem like a very insignificant amount of building blocks to compute anything..

Lastly, how would/will qubits be used to revolutionize storage? I get the allure of storing bits on a subatomic level but if the whole hype is about storage density, it sort of kills the magic for me.

Re:Explain the hype, please?

[noundi]

Ok, so on this site bursting with intelligent, educated folk...

You lost me at "Ok".

Re:Explain the hype, please?

[maxwell demon]

Ok, I just wrote a lengthy reply, and then by accident hit "refresh", and all the text was gone :-(

Therefore here the short version:

• The speedup is basically because for quantum systems the dimension of the configuration space grows exponentially rather than linearly with the size of the system (i.e. number of qubits). The fact that we can't simply measure the complete state is actually a limitation, because it means we cannot directly access an arbitrary unknown state.
• You can do quantum computing by just doing measurements because every measurement modifies the measured system, and with entangled states, this change is non-local (i.e. you also modify parts of the system where you didn't just destroy any entanglement by your measurement). However you need special entangled states to do universal measurement-based quantum computing (i.e. to allow arbitrary transformations with measurement only); one state which works is the cluster state produced by this "photon machine gun"
• They didn't claim that qubits revolutionize storage, but that if emulating the 20 to 30 qubit quantum computer on a classical computer, it would not fit into computer storage. However I doubt that; storing the state of 30 qubits needs about 16 GB, which is large, but perfectly doable in todays computers (and may be actually standard by the time this photon gun is realized). The problem with simulating the quantum computer would not be storage, but time.

Re:Explain the hype, please?

[Anonymous Coward]

Bad analogy time.

The simplest way to factor a large number is to just try to divide it by 2, by 3, etc. Once you've divided it, you now have 2 smaller numbers to factor. Repeat until you get a prime. This takes a long time for a large number because you have to try it over and over again.

With a quantum computer you can do all of these computations in parallel, and then arrange for all of the non-factors to cancel each other out, meaning that you can only measure a legitimate factoring. (Getting all of the non-answers to cancel out is the trick in quantum computing, and it isn't a particularly easy one to pull off. These are not general purpose computers!) If it keeps giving you 1*n as your factoring, eventually you conclude it is prime. Otherwise the first time it gives you something else, you've broken the number down into 2 easier ones.

To break RSA you only have to factor one number. So everyone cites that as the classic problem. But you can't factor a number you can't put in. With 20-30 qbits you can only input 20-30 bit numbers so you can't factor anything bigger than that. By contrast a motivated person these days with a few PCs and a few months to devote to it can factor a general 600 bit number. Most people's codes are 1024 bits or longer.

Therefore this research is cool, but any claim of an immediate threat to cryptography is waaay overblown.

Re:Explain the hype, please?

[noundi]

Ok, I just wrote a lengthy reply, and then by accident hit "refresh", and all the text was gone :-(

You're welcome.

Re:Qubit does not double power in traditional sens

[MartinSchou]

If you are very unfortunate, n qubits can map 2^n -1 bits. -1 because 2^0 = 1, and that'd just be weird.

If this is the case, then a 6 qubit machine maps 63 bits, but 20 would map 1,048,575 bits (1 Mbit of information) and 30 would map 1 Gbit of information.

Dirty Erwin

[ciderVisor]

I know what you're thinking: "Did he flip six qbits or only five?" Well, to tell you the truth, in all this excitement I kind of lost track myself. But being as this is a Photon Machine Gun, the most powerful quantum entanglement source in the world, and would blow your head clean off, you've got to ask yourself one question: Is the cat dead or alive ? Well, is it, punk ?

Heh

[L4t3r4lu5]

The quantum machine gun is described as 'one of the most exciting theoretical proposals I've read in five years' by a leading quantum physicist.

The long winter nights must just fly by.

How Shor's algorithm works (the problem with QC)

[professorguy]

The quantum algorithm for factoring does not just divide repeatedly "in parallel." Shor's algorithm really describes a specially built machine for factoring (which converts factoring to period finding, and a fourier analysis is forced and sampled).

In fact, I first studied Shor's algorithm in order to understand why good programmers weren't looking at it, generalizing it, and writing a million more algorithms. I was disappointed to learn that we are not far enough along to describe a universal quantum computer and are still in the mode of building special-purpose machines. Like the early bomb dropping "computers" of WW2, results are still being generated but without a concept of universality.

We don't have the math for universal quantum computation since it is still unknown what they are capable of (what their universe consists of). Until the math arrives, we're stuck with this scatter-shot approach.

Re:Qubit does not double power in traditional sens

[Anonymous Coward]

A prime number is prime in any base. You can't gain any complexity by using another base like you suggest.