https://www.newscientist.com/articl...-googles-plan-for-quantum-computer-supremacy/
More in the article! Interesting stuff, and when this is eventually achieved, it will open up a lot of very unique opportunities. Hybrid classic/quantum computing systems would be super cool.
This is a great article from last year that goes over a lot of stuff, and talks more about the Martinis hire, as well as machine learning https://www.technologyreview.com/s/544421/googles-quantum-dream-machine/
The machine learning specific excerpt:
SOMEWHERE in California, Google is building a device that will usher in a new era for computing. Its a quantum computer, the largest ever made, designed to prove once and for all that machines exploiting exotic physics can outperform the worlds top supercomputers.
And New Scientist has learned it could be ready sooner than anyone expected perhaps even by the end of next year.
...
The firms plans are secretive, and Google declined to comment for this article. But researchers contacted by New Scientist all believe it is on the cusp of a breakthrough, following presentations at conferences and private meetings.
"They are definitely the world leaders now, there is no doubt about it, says Simon Devitt at the RIKEN Center for Emergent Matter Science in Japan. Its Googles to lose. If Googles not the group that does it, then something has gone wrong.
We have had a glimpse of Googles intentions. Last month, its engineers quietly published a paper detailing their plans (arxiv.org/abs/1608.00263). Their goal, audaciously named quantum supremacy, is to build the first quantum computer capable of performing a task no classical computer can.
...
To push classical computing to the limit, Google turned to Edison, one of the most advanced supercomputers in the world, housed at the US National Energy Research Scientific Computing Center. Google had it simulate the behaviour of quantum circuits on increasingly larger grids of qubits, up to a 6 × 7 grid of 42 qubits.
This computation is difficult because as the grid size increases, the amount of memory needed to store everything balloons rapidly. A 6 × 4 grid needed just 268 megabytes, less than found in your average smartphone. The 6 × 7 grid demanded 70 terabytes, roughly 10,000 times that of a high-end PC.
Google stopped there because going to the next size up is currently impossible: a 48-qubit grid would require 2.252 petabytes of memory, almost double that of the top supercomputer in the world. If Google can solve the problem with a 50-qubit quantum computer, it will have beaten every other computer in existence.
...
Google purchased its D-Wave computer in 2013 to figure out whether it could be used to improve search results and artificial intelligence. The following year, the firm hired John Martinis at the University of California, Santa Barbara, to design its own superconducting qubits. His qubits are way higher quality, says Aaronson.
Its Martinis and colleagues who are now attempting to achieve quantum supremacy with 50 qubits, and many believe they will get there soon. I think this is achievable within two or three years, says Matthias Troyer at the Swiss Federal Institute of Technology in Zurich. Theyve showed concrete steps on how they will do it.
More in the article! Interesting stuff, and when this is eventually achieved, it will open up a lot of very unique opportunities. Hybrid classic/quantum computing systems would be super cool.
This is a great article from last year that goes over a lot of stuff, and talks more about the Martinis hire, as well as machine learning https://www.technologyreview.com/s/544421/googles-quantum-dream-machine/
The machine learning specific excerpt:
When Martinis explains why his technology is needed at Google, he doesnt spare the feelings of the people working on AI. Machine-learning algorithms are really kind of stupid, he says, with a hint of wonder in his voice. They need so many examples to learn.
Indeed, the machine learning used by Google and other computing companies is pathetic next to the way humans or animals pick up new skills or knowledge. Teaching a piece of software new tricks, such as how to recognize cars and cats in photos, generally requires thousands or millions of carefully curated and labeled examples. Although a technique called deep learning has recently produced striking advances in the accuracy with which software can learn to interpret images and speech, more complex faculties like understanding the nuances of language remain out of machines reach.
Figuring out how Martiniss chips can make Googles software less stupid falls to Neven. He thinks that the prodigious power of qubits will narrow the gap between machine learning and biological learningand remake the field of artificial intelligence. Machine learning will be transformed into quantum learning, he says. That could mean software that can learn from messier data, or from less data, or even without explicit instruction. For instance, Googles researchers have designed an algorithm they think could allow machine-learning software to pick up a new trick even if as much as half the example data its given is incorrectly labeled. Neven muses that this kind of computational muscle could be the key to giving computers capabilities today limited to humans. People talk about whether we can make creative machinesthe most creative systems we can build will be quantum AI systems, he says.
...
Neven is confident that Googles quantum craftsmen and his team can get through all that. He pictures rows of superconducting chips lined up in data centers for Google engineers to access over the Internet relatively soon. I would predict that in 10 years theres nothing but quantum machine learningyou dont do the conventional way anymore, he says. A smiling Martinis warily accepts that vision. I like that, but its hard, he says. He can say that, but I have to build it.