DailyDirt: Quantum Computing Works Now (Sorta)
from the urls-we-dig-up dept
If you haven’t heard about quantum computing, it’s an alternative to “classical computing” that relies on some strange properties of quantum physics. Sure, the computer (or whatever device) you’re reading this on also relies on physics a bit, but it stores information digitally as ones and zeroes — and not some superposition of two states of matter in an array. A few organizations are working on quantum computers (e.g., Google, NASA, D-Wave, Cambridge Quantum Computing, Yale Quantum Institute, Microsoft, IBM, etc.), but the true potential is still just slightly out of reach (for now).
- A D-Wave quantum computer tested by Google (and hosted by NASA) has been demonstrated to outperform a traditional computer — at least for one very specific task. Google has been trying to see what it can do with more and more qubits, and it looks like if Google and D-Wave can keep adding more qubits to their system, they might actually get a quantum computer that can beat our old digital computers (without rigging up a contrived test). [url]
- Cambridge Quantum Computing is working on a quantum computer operating system. It’s a bit difficult to imagine how an operating system is created when the underlying hardware can potentially be built in numerous (and sometimes unproven) ways. [url]
- Curious folks can play with a virtual 22-qubit quantum computer — the Quantum Computing Playground (from Google). This demo runs in a browser and just simulates how a quantum computer could work with a scripting language. This isn’t going to print “Hello, World!” a gazillion times. [url]
- If you’re wondering what quantum computers might be good for, one of the examples is “optimization problems” — such as a class of math problems that are NP hard, like the traveling salesman problem. Some folks think quantum computing will unlock a breakthrough for artificial intelligence, allowing computers to “brute force” their way into tackling complex problems without having to invent faster and smaller semiconducting chips. (But there’s a chance that we’ll have to figure out how to invent ways to make more and more qubits?) [url]
After you’ve finished checking out those links, take a look at our holiday gift ideas for cool gadgets and other awesome stuff.
Filed Under: np-hard, operating system, optimization, quantum computers, quantum computing playground, quantum physics, qubits, shor's algorithm
Companies: cambridge quantum computing, d-wave, google, ibm, microsoft, nasa, yale quantum institute
Comments on “DailyDirt: Quantum Computing Works Now (Sorta)”
I Think These Are Just Analog Computers
Up to around the middle part of the 20th century, when digital electronics were much slower, there were “analog” computers. These implemented electronic analogs of the physical models they were trying to solve. While being much faster than the digital computers of the time, they also gave much less accuracy (maybe 3-4 significant figures).
I have a feeling the D-Wave is the same sort of thing. It may be 100 million times than conventional digital electronics, but I don’t think it can match the accuracy of the latter.
Re: I Think These Are Just Analog Computers
YOu are right – although of course one has to be clear what analog actually means – it DOESN’T mean “continuous” it means “by analogy” as you sort of point out.
To talk about quantum computers “outperforming” classical computers is of course nonsense. Quantum computers are horrendously difficult to “program” effectively. The entire research community has managed to create a set of algorithms that you can count on the fingers of one hand in the last 20 years.
What they can do is provide effective simulation for quantum mechanical systems.
For some reliable – hype free – information look at Scott Aaronson’s blog – here:
Only one of the “traveling salesmen” is real; any others are imaginary (more like fantasy than virtual). Quantum computing, per se, may turn out to be just as imaginary… but that’s science for ya.
Re: Travelling Salesman
Actually the idea that quantum computing can solve the travelling salesman or other similar problems is a myth mostly propagated by journalists and others with a poor understanding of the subject. Some QC researcher cynically avoid undermining this myth because the myth is good for their funding!
Famously QC CAN factor big numbers very fast – but that is only because factorisation is NOT one of the general set of problems to which the travelling salesman belongs. Factoring has more structure than NP complete problems and hence is “easier”. It is this structure that Shor’s algorithm exploits. So I’d say that the final link in the article is misleading rubbish.
Re: Re: Travelling Salesman
Well, there are NP-complete problems and NP-hard problems. NP-complete means (simplified version) that it’s very difficult to compute, but easy to verify the right answer once you see it. Factorization belongs to this category: once you have the factors, it’s simple arithmetic to verify that they are indeed the factors of the number in question.
The Traveling Salesman problem, on the other hand, is not easy to verify. It asks which permutation of possible routes the salesman may take is the fastest, and this necessarily requires you to compute every possible permutation of routes (or at least a significant percentage of them, if you’re clever enough to find a way to prove some of them can be pruned early) and then see which of them is the shortest. A problem that’s very difficult to both compute and verify is called NP-hard, and they’re significantly worse than ordinary NP-complete problems.
Re: Re: Re: NP-complete means (simplified version)
“NP-complete” has to do with the “P = NP” conjecture. NP-complete problems are NP-hard; they just have the additional property that, if any of them can be proven to be of the easier class P (solvable in deterministic polynomial time) instead of NP (solvable in non-deterministic polynomial time), then there is no such thing as a separate class NP, all such problems actually being in class P.
Re: Re: Re:2 NP-complete means (simplified version)
Yes, I know, but that was a deliberately simplified post for the benefit of people without a background in computer science and complexity theory. 😛
Re: Re: Re: Travelling Salesman
NP-complete means (simplified version) that it’s very difficult to compute, but easy to verify the right answer once you see it. Factorization belongs to this category:
No it doesn’t! In this diagram – assuming as most do that the left side is correct:
then factoring is in NP but not in NP complete. You can’t use Shor’s algorithm to solve 3sat.
We know you typed the word “yes” in Word. And we’re telling you, there’s a 50% chance it will come out “no”. That’s just Microsoft Windows works at the quantum level. Yes, the upgrade is mandatory. (Or maybe no?)
I have to wonder where this is going...
This could be just dystopian fantasy, but somehow I suspect that one of the kinds of problems quantum computing will be set to solve will be cracking encryption that is presently computationally unfeasible to break.
Possibly we’re seeing the first steps toward what will be the end of privacy and security based on mathematics.
Re: I have to wonder where this is going...
Anything that has enough power to decrypted keys can create keys as strong, and thanks to data speeds a bigger key aint too much of an issue 🙂
Re: Re: I have to wonder where this is going...
Actually not true – but quantum encryption – which is closer to feasible than quantum factoring – also fixes the problem.
Re: Re: I have to wonder where this is going...
I don’t think this is true. Post-quantum cryptography generally involves different algorithms, not just larger keys. Some of the algorithms do have huge keys/signatures, but you can’t just switch to a huge RSA key to get security (and I don’t know of any reason why a quantum computer would be better at modular exponentiation than a classical one).
Re: Re: Re: I have to wonder where this is going...
Thank you for the post-quantum cryptography link. I see they’re way ahead of me.
Looking forward to the quantum cat vids – i can haz existenz?
Fully working quantum computers already exist. I saw one on the show Scorpion. It’s based on a real person’s life, they wouldn’t just make something like that up…
Like I'm a billionaire...