Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
It’s been almost two decades since Peter Shor came up with a a breakthrough algorithm for finding the prime factors of a number with a quantum computer, sparking great interest in quantum computing. But commercial adoption has been pretty much nonexistent. On Thursday, though, Google (s goog) came forward with news that it’s launching a Quantum Artificial Intelligence Lab that will include a quantum computer, apparently making it the second company to pay for a quantum computer. The development suggests that quantum computing could finally be taking off.
Earlier this year Lockheed Martin shared details of its implementation of a D-Wave Systems quantum computer, which reportedly cost $10 million: The contractor is using the computer to develop new aircraft, radar and space systems.
Now Google is taking steps at incorporating more quantum computing into its operations with the Quantum Artificial Intelligence Lab, which will be located at the NASA Ames Research Center in Moffett Field, Calif. Researchers from the Universities Space Research Association will be able to use the machine 20 percent of the time, Forbes reports. That could lead to lots of interdisciplinary thinking and collaboration.
For Google, though, the goal of the initiative is to make strides in machine learning, according to a Thursday Google Research blog post. The best results could trickle down to end users, perhaps in search results and speech-recognition applications.
Quantum computing could mean smarter smartphones
Google has already assembled machine-learning algorithms that involve quantum elements, Hartmut Neven, a Google director of engineering, explained in the post:
One produces very compact, efficient recognizers — very useful when you’re short on power, as on a mobile device. Another can handle highly polluted training data, where a high percentage of the examples are mislabeled, as they often are in the real world.
It’s not hard to imagine how quantum computing could inform machine learning on a smartphone with just a drop of battery life left. It could be that a smarter smartphone one day will take a minuscule amount of input and determine with a high probability who a user wants to talk to or what information it needs right away, rather than forcing the user to cycle through a string of commands and risking the death of the battery altogether.
The applications might have arisen after Google’s earlier partnership with D-Wave, which came to light in a different blog post from Neven in 2009.
Google has already used machine learning to recognize faces and other things in photos and videos. New technology Google executives talked about at the Google I/O developer conference in San Francisco on Wednesday also appears to use machine learning to stitch together photos and clean them up.
What Google has learned so far is the best results come from blending regular binary computing using ones and zeros with quantum style computing. Quantum computing accommodates the space between a one and a zero with quantum bits of information, or qubits. It can express likelihood as well as take shortcuts by approximating when handling certain kinds of workloads. Given what Google has observed thus far, it could decide to build hardware combining quantum and classical computing capabilities.
For now, though, Google is diving deeper into quantum computing with the D-Wave machine. The move could kick off a sort of arms race for webscale companies to buy quantum computers and come up with new notions by way of probabilistic logic. In this way, Google could help push the development of quantum computing much like its invention of MapReduce changed the way firms do distributed data processing.
In any case, quantum computing has a long way to go before reaching commercial viability. That could take decades (so far it has). But because the organization at the helm of the quantum research is Google and not IBM (s ibm) or Bell Labs, regular people could start seeing much more of the advantages in just a few years’ time, which in turn could drive commercialization.
Feature image courtesy of Shutterstock user pixeldreams.eu.