May 16, 2013 at 8:41 PM ET
Google isn't satisfied with dominating just ordinary computers, it seems: The Internet giant is making a big investment in quantum computing with a new lab and partnership. The company hopes to develop applications for the strange but powerful technology.
The Quantum Artificial Intelligence Lab, a cooperative effort with NASA and the Universities Space Research Association, will investigate "how quantum computing might advance machine learning," according to a post on the Google Research blog.
What that means is a bit difficult to explain. Machine learning isn't just about the Terminators and A.I.'s of science fiction, but describes a whole set of computing problems that are not effectively solved by the brute-force processing power of traditional computing.
Classical computers, which use networks of millions of tiny electrical switches (transistors) to perform mathematical calculations, rely on ones and zeroes — bits — as their most basic elements. But quantum computers operate on completely different and more flexible rules. Instead of binary bits, they use quantum bits, or "qubits," which can represent ones and zeroes at the same time. It's no simple task to make them work, but they excel in some areas that digital computers fail.
Google uses the basic example of finding the lowest point of ground on a virtual surface of hills and valleys. Traditional computing starts at a spot and searches progressively and systematically: Is it here? How about here? This can be a slow process, and unless you have unlimited time and patience, you may have to be satisfied with a "good-enough" answer — after all, a new, deeper valley might turn up after a few more calculations.
Quantum computers could conduct a search of that sort much more efficiently, because they could consider multiple values simultaneously. The goal is to achieve massive improvements in computing performance.
How massive? Colin Williams of D-Wave, the company that makes the quantum devices to be used in the new lab, told The New York Times that for a few complex problems, its device was 50,000 times faster. Such claims are based on extremely special conditions, and must be taken with a grain of salt. But if the company has a quantum device that works as well as it says, that's an accomplishment in and of itself.
It won't make a difference when you're sending an email or listening to music on your PC or phone. But for certain complex tasks, it could bring major changes. As Google Research's Director of Engineering, Hartmut Neven, writes in the blog post:
We hope it helps researchers construct more efficient and more accurate models for everything from speech recognition, to web search, to protein folding. We actually think quantum machine learning may provide the most creative problem-solving process under the known laws of physics.
That said, it isn't likely to kill off digital computing anytime soon: Quantum devices have their own limitations, not least of which is that they're rather hard to make, and incredibly expensive.
At this point, it's all still mostly theoretical, which is why the Quantum Artificial Intelligence Lab is being formed. The goal is to apply the theory to real problems, and write real code for real quantum devices. Google declined to comment further on the plans or personnel involved.
They're not the only ones in the field: Microsoft Research's Station Q is looking at quantum-computing solutions for complex mathematical problems, and plenty of other research organizations are working on related topics. But this Google-NASA venture will still find itself in fairly exclusive company.
Devin Coldewey is a contributing writer for NBC News Digital. His personal website is coldewey.cc.