Quantum Computer Made From Photons Achieves New Record (scientificamerican.com) 23
Slashdot reader hackingbear shared this article from Scientific American:
In the race to create a quantum computer that can outperform a classical one, a method using particles of light (photons) has taken a promising step forward. Jian-Wei Pan and Chao-Yang Lu, both at the University of Science and Technology of China, and their colleagues improved a quantum computing technique called boson sampling to achieve a record 14 detected photons in its final results. Previous experiments were capped at only five detected photons. The increase in the number of the particles is small, but it amounts to a 6.5-billion-fold gain in "state space," or the number of ways in which a computer system can be configured. The larger the state space, the less likely a classical computer can perform the same calculation.
The result was reported in a paper posted at the preprint server arXiv.org on October 22 and has yet to be peer-reviewed. But if it is confirmed, it would be an important milestone in the race for quantum-computational supremacy -- a fuzzy goalpost defined as the point where quantum computers outpace their best classical counterparts.... Pan and Lu argue in their paper that their technique is another possible route toward quantum supremacy... Part of the trouble is its limited utility. "A universal computer can solve any different type of problem," says Jonathan Dowling, a theoretical physicist at Louisiana State University, who was not involved with the research. "This can only solve one." But solving just one problem faster than a classical computer would count as a demonstration of quantum-computational supremacy...
Over the past few weeks, the race for quantum computational supremacy has reached a breakneck pace. Google's quantum computer performed an operation that its scientists claim would take a classical computer 10,000 years in just 200 seconds. IBM researchers, who are also working on a quantum computer, have expressed doubts, suggesting a classical computer could solve that problem in under three days... "Quantum supremacy is like a horse race where you don't know how fast your horse is, you don't know how fast anybody else's horse is, and some of the horses are goats," Jonathan Dowling, a theoretical physicist at Louisiana State University, says. But this result, he clarifies, is not a goat.
The result was reported in a paper posted at the preprint server arXiv.org on October 22 and has yet to be peer-reviewed. But if it is confirmed, it would be an important milestone in the race for quantum-computational supremacy -- a fuzzy goalpost defined as the point where quantum computers outpace their best classical counterparts.... Pan and Lu argue in their paper that their technique is another possible route toward quantum supremacy... Part of the trouble is its limited utility. "A universal computer can solve any different type of problem," says Jonathan Dowling, a theoretical physicist at Louisiana State University, who was not involved with the research. "This can only solve one." But solving just one problem faster than a classical computer would count as a demonstration of quantum-computational supremacy...
Over the past few weeks, the race for quantum computational supremacy has reached a breakneck pace. Google's quantum computer performed an operation that its scientists claim would take a classical computer 10,000 years in just 200 seconds. IBM researchers, who are also working on a quantum computer, have expressed doubts, suggesting a classical computer could solve that problem in under three days... "Quantum supremacy is like a horse race where you don't know how fast your horse is, you don't know how fast anybody else's horse is, and some of the horses are goats," Jonathan Dowling, a theoretical physicist at Louisiana State University, says. But this result, he clarifies, is not a goat.
all this quantum computing leaves me feeling fuzzy (Score:2)
it looks good in theory, but in practical terms, it's going to be a while before QC affects me.
640 qubits... (Score:3)
...ought to be enough for anybody.
Re: (Score:2)
in practical terms, it's going to be a while before QC affects me.
Yeah, it'll be a handful of decades before we have anything interesting.
Re: (Score:2)
The idea under discussion (Score:4, Interesting)
The idea under discussion, quantum supremacy, is a bit subtle. Essentially the idea is that getting quantum computers to do a lot of the things we want quantum computers to be able to do that are apparently tough for classical computers, like factoring large numbers, requires large numbers of high quality qubits. Quantum supremacy instead focuses on just getting a quantum computer to do some sort of computation that we know is hard for a classical computer. This has two advantages. First, it doesn't require nearly as much fine tuned cooperation of qubits, and second this can then rely on much weaker computational complexity conjectures, which are substantially less weak than factoring being hard, and are things like the polynomial hierarchy not collapsing https://en.wikipedia.org/wiki/Polynomial_hierarchy [wikipedia.org]
It turns out that sampling problems, where one cares about distributions rather than specific calculations are a natural context for quantum supremacy. In this particular case, the particular class of problems is known as boson sampling https://en.wikipedia.org/wiki/Boson_sampling [wikipedia.org] due to Scott Aaronson and Alex Arkhipov. The idea is essentially that all one has is a bunch of photons with beam splitters and a few other very simple gadgets, but no actual interaction between photons, and one wants to know the output distribution. They showed that under very weak assumptions (not as weak as P != NP but only a tiny bit stronger), a classical computer cannot efficiently do this process. So if one can show that one really can do boson sampling efficiently one will have shown that quantum computers really can do at least some things that a classical computer cannot do efficiently.
Note that until very recently it looked like boson sampling was essentially out of the running for the first unambiguous quantum supremacy demonstration because of Google's recent work with random general quantum circuits. But this work seems to potentially put boson sampling back in as a serious contender.
Re: (Score:2)
The beam splitters and simple gadgets could solve the problem a whole lot cheaper, so it's more like quantum inferiority.
I see it as just an overly complex way of reproducing classical interference for the moment.
Re: (Score:2)
Also quite interesting is that someone tagged the story "nonsense" and "useless". Either they are hard-core quantum sceptics or it's because the research is coming from China.
Slashdot is slowly getting worse.
Universal computer (Score:2)
Is there any useful comparison between current quantum computers and a Turing machine? What is/are the missing elements? Can quantum computers even store and modify intermediate state at all?
Re: (Score:2)
Is there any useful comparison between current quantum computers and a Turing machine?
Nope. The main difference is that elements of a Turing machine are independent component that interact via a simple interface. A QC needs to be _one_ component or it does not work. So while you can actually do everything with a QC that a Turing machine can do, it does make zero sense to do so, as QCs are extremely inefficient at most operations and scale so badly that they will very likely never be of any use in this universe.
Odds of getting this on a chip within 20 years? (Score:3)
Quantum computing sure is interesting as an intellectual exercise.
Does anyone who knows about the processes of making semiconductor chips have any insight on how things look for putting thousands of qubits on a chip using something resembling current semiconductor processes? On a practical level, things get real interesting when many organizations have quantum computers with thousands or millions of cubits.
Are any of the quantum tech ideas that are advancing stuff that can be done in silicon, or is this going to be 100 pounds per qubit for the foreseeable future?
Re: (Score:2)
Putting the bits in there is trivial. Getting and keeping them entangled is not. And without the second, the whole thing is completely useless as a computing mechanism.
Re: (Score:2)
Getting from bits is likely a quicker route to faster computer. Transistors that can hold more than just two states, more holding a frequency rather than just an on off energy state, say capable of holding ten different frequencies, a decimal computer rather than a binary computer. Processing far more data at the same time, more complex instructions that can be calculated simultaneously.
My computer is made out of electrons (Score:2)
Exclusively, according to the stupid title.
Re: (Score:2)
You own a quantum computer! Kudos.
peer review (Score:2)
"...and has yet to be peer-reviewed. But if it is confirmed, it would be an..."
Peer review does not confirm anything. It is essentially spam filtering + some basic quality and completeness checks. A finding is confirmed when it is reproduced by other groups. The most important function of peer reviewers is to check that the method is written down in a way so that it can be reproduced by others.
Re: (Score:2)
I do agree that the sentence structure was not as good as it could have been.
Is this guy a /.er? (Score:3)
"Quantum supremacy is like a horse race where you don't know how fast your horse is, you don't know how fast anybody else's horse is, and some of the horses are goats,"