Slashdot Mirror
China Makes Quantum Leap In Developing Quantum Computer (scmp.com)
hackingbear writes: Researchers at the University of Science and Technology of China created a quantum device, called a boson sampling machine, that can now carry out calculations for five photons, but at a speed 24,000 times faster than previous experiments. Pan Jianwei, the lead scientist on the project, said that though their device was already (only) 10 to 11 times faster at carrying out the calculations than the first electronic digital computer, ENIAC, and the first transistor computer, TRADIC, in running the classical algorithm, their machine would eclipse all of the world's supercomputers in a few years. "Our architecture is feasible to be scaled up to a larger number of photons and with a higher rate to race against increasingly advanced classical computers," they said in the research paper published in Nature Photonics. This device is said to be the first quantum computer beating a real electronic classical computer in practice. Scientists estimate that the current faster supercomputers would struggle to estimate the behavior of 20 photons.
I always struggle with understanding quantum computing concepts, but from the sound of things in the article, this is not some sort of general purpose quantum computer. Rather, it's a purpose-built computer dedicated to estimating the behavior of photons.
Why that specifically?
Based on what the article (and summary) said, modern computers struggle to estimate the behavior of 20 or more photons, but it's the sort of problem that quantum computers are theoretically capable of handling quite easily. Researchers are apparently suggesting that in order to disprove skeptics and bring in more support for quantum computing, we should build a quantum computer of this variety and then use it to estimate the behavior of 30 or more photons, because doing so would definitively prove to everyone that quantum computers can provide a massive advantage over traditional computing methods.
Nope, I mean to the scientific paper in Nature Photonics, not press-release...
Like this: http://www.nature.com/nphoton/...
Paul B.
It is both a link to the article and a link not to the article. Being quantum physics, you'll never know which it is until you click it.
Paper preprint...
Wikipage about boson sampling...
In principle, a large-scale boson-sampling machine would constitute an effective disproof against a foundational tenet in computer science: the Extended Church-Turing Thesis, which postulates that all realistic physical systems can be efficiently simulated with a (classical) probabilistic Turing machine.
The machine may not have any practical use, but it still is an interesting theoretical advance that might serve to challenge our understanding of computablity... Part of the theoretical importance of this area of research is the understanding of #P-complete problems.
The wikipedia articlenotes the theoretical significance of this...
A polynomial-time algorithm for solving a #P-complete problem, if it existed, would imply P = NP, and thus P = PH. No such algorithm is currently known.
Scott Bakula could not be reached for comment.
I doubt the headline writer saw what they did there though. A quantum leap is literally the smallest possible change to a system. So the headline suggests they have made the smallest possible improvement which is not very impressive at all.
A quantum leap is hardly noteworthy. Literally that is the smallest possible motion, used in physics for the smallest leaps within an atom.
True, but a quantum leap can occur over an energy-boundary that classical physics would claim can't be overcome. I think that's why the metaphor is applied frequently to an unexpected advance in various fields outside of quantum mechanics.
If it weren't for deadlines, nothing would be late.
You set a register of bits to all possible combinations of the bits at the same time -- all possible values from 0 to N^2-1 are entangled. Then you run them through some quantum logic operations that eliminate all impossible solutions to a problem from the set of possible combinations. Then you read the register. It collapses to *one* of the possible solutions when you read it. So for example if you are factoring a large umber, that solution will be one possible divisor of that number. So you divide the large number by that number, then use the same process to factor those two smaller numbers. Where you save time is you did not have to iterate through a large bunch of prime numbers that were not factors by trying to divide the large number by each of them in turn.
As far as detecting particles, the "probabilities" are better when you only need a binary result like "spin up versus spin down". So even if you have a probability of erroneous readings, some of the readings will be good, and when you are looking for needles in a haystack, as long as you can tell when you found a needle, all you need is a significantly better probability of a good reading than the probability of randomly sticking your hand in and finding the needle.
Someone had to do it.
I suggest bigger cuts to the Office of Science budget. Why do we need to spend money developing better, faster supercomputers? We can let the Chinese do all the expensive R&D, then we can buy the finished product from them. No problem. It worked for drywall, why not quantum puters?
The Russians have won. They have made the world a cesspool of distrust, greed, fear and hate.
So their breakthrough is a vanishingly small one?