Breakthrough Toward Quantum Computing

redwolfe7707 writes "Qubit registers have been a hard thing to construct; this looks to be a substantial advance in the multiple entanglements required for their use. Quoting: 'Olivier Pfister, a professor of physics in the University of Virginia's College of Arts & Sciences, has just published findings in the journal Physical Review Letters demonstrating a breakthrough in the creation of massive numbers of entangled qubits, more precisely a multilevel variant thereof called Qmodes. ... Pfister and researchers in his lab used sophisticated lasers to engineer 15 groups of four entangled Qmodes each, for a total of 60 measurable Qmodes, the most ever created. They believe they may have created as many as 150 groups, or 600 Qmodes, but could measure only 60 with the techniques they used.'" In related news, research published in the New Journal of Physics (abstract) shows "how quantum and classical data can be interlaced in a real-world fiber optics network, taking a step toward distributing quantum information to the home, and with it a quantum internet."

Re:um...

[jmizrahi]

Neither statement is true. First, we have entangled many systems other than photons. We have entangled trapped ions, neutral Rydberg atoms, superconducting qubits, nuclear spin states, and the list goes on. There are advantages and disadvantages to each quantum computing architecture. One of the fundamental issues facing all quantum computing architectures is the question of scalability. It is not always clear how to go from 1 or 2 qubits to thousands or millions of qubits. Some architectures, such as trapped ions, lend themselves naturally to scaling. The significance of this work is that up to this point, it has been unclear how you might scale a photonic quantum computer. The authors of this paper have taken the first steps towards overcoming that obstacle. As to your second statement, observed photon entanglement cannot be explained via classical optics. It has been shown to violate a Bell inequality, which is the hallmark of non-classicality in quantum mechanics.

Re:How do they work?

[Alsee]

We can compare it to rolling dice, where the dice can be loaded to shift the percentages. If we put a weight on the 1, we might roll a 1 half the time and randomly get 2 through 6 the rest of the time. In quantum mechanics we can preform calculations that change how the dice are loaded. Ideally, we can load the die so strongly that 2 through 6 are driven down to zero percent, and 100% of the time we "randomly" roll a 1.

Depending on the particular problem and the particular technology used, certain parts of the computer might not be working with a perfect-clean 100%. Particular parts of the computer might have the dice loaded to 99.9% randomly roll a particular result. Obviously we don't want a computer that's oly 99.9% right :D

Different kinds of quantum computers deal with that in different ways. The simplest example is a quantum computer that works on a beam of photons or something. A beam of light might contain a trillion photons per nanosecond. If 99.9% of those photons randomly come out on the right answer, the right answer obviously lights up brightly. The 0.1% of photons lighting up the various wrong answers will be too dim to notice.

For some quantum calculations they use the very simple technique of just running it a dozen times or something. There's basically a 99% chance you'll get a dozen matching correct answers, and a 1% chance you'll get eleven matching correct answers and one random error that you throw away. There is a minuscule chance you'll get (and throw away) two or possibly three garbage results out of the dozen, but the only way you would ever get a wrong answer from that is if the same wrong answer came up seven or more times at once, and mathematically that won't happen a million or billion years.

However for a "real" desktop-type quantum computer, there's a much more complicated and powerful technique they would build in.... error correcting bits and error correcting codes. By adding in a few extra bits, the computer can automatically spot and correct any random wrong values as soon as the appear. All of the automatic error correction stuff might make the computer something like 50% bigger or maybe twice the size, but it can easily match the (effectively zero) error rate of standard computers.

-