Quantum computing? Yes, No and Maybe
Canadian company D-Wave has sold its machine to Google and NASA but question marks remain over how it works. Bill Condie reports.
Quantum computers speak the language of probability and chance, a far cry from the straight-talking binary language of today’s computers. But their fuzzy logic promises to tease out answers to problems that are currently impossible even for the world’s largest supercomputers – like cracking encryption codes, searching gargantuan databases and modelling proteins.
Problem is, despite claims to the contrary, many scientists think they are light years away from delivering. But recent events have forced even the most hard-nosed sceptics to take another look. Canadian company D-Wave claims to produce the world’s first commercial quantum computer and in May was given a boost when NASA and Google ordered one of its machines at a cost of US$15 million. They will co-operate on its use.
D-Wave’s first customer was Global aeronautical engineering and defence giant Lockheed Martin, which proposed using it to solve “some of Lockheed Martin’s most challenging computation problems”. Chemists at Harvard University also teamed up with D-Wave researchers to solve, at least partially, the problem of how best to fold the amino acids of a protein chain so that it takes on the most stable structure – research published in the journal Scientific Reports in August, 2012.
The D-Wave machine bears more than a passing resemblance to the 'Infinite Improbability Drive'.
The most recent buyers are also counting on D-Wave to do some fuzzy problem solving. Google hopes to make our internet searching easier and more intuitive, while NASA thinks it will help it find Earth-like planets by streamlining searches through mountains of data on objects that may or may not be planets.
The D-Wave machine bears more than a passing resemblance to the “Infinite Improbability Drive” that Douglas Adams fantasised in Hitchhikers’ Guide to the Galaxy. And indeed, notwithstanding the corporations’ buy-in, the jury is still out as to whether D-Wave’s claims are science or science fiction.
But the concept of a quantum computer is quite real, and there is no doubt progress is being made.
The ultimate quantum computer would use truly quantum particles, a single atom or smaller. It would not calculate in black and white the way a classic computer does, 1 or 0; yes or no. In the mysterious quantum world, particles have a property called “spin”, equivalent to magnetic orientation. Their spin state can either be “up”, “down” or, weirdly, both at once – a so-called “superposition”. So a quantum bit, a “qubit”, can be 1 or 0, or both 1 and 0; yes, no and maybe. Somehow this allows the machine to simultaneously try multiple possible solutions to an astronomically difficult problem, in theory finding the answer much more quickly than a classical binary computer.
One outfit whose claims invite no scepticism is Andrea Morello and his team at the University of New South Wales in Australia. They hold the distinction of being the first in the world to create a working qubit based on a single atom of phosphorous in a silicon microchip. In three landmark papers published in Nature magazine in 2010, 2012,and 2013, they showed it was possible to read, write and retrieve information from their qubit.
Morello saw the D-Wave machine up close in June this year in the company’s Vancouver lab. It was not enough to turn him into a convert but it did satisfy his curiosity about its construction and left him impressed by the engineering.
At the core of the machine is a large electrical circuit and the D-Wave qubit – not a particle like Morello’s but a loop of superconducting material. Its spin state comes from switching the magnetic field based on the directional flow of the current. “You can think of it as an artificial spin, where the spin up or down is the current flowing clockwise or counter-clockwise in this circuit,” Morello explains.
One problem for the machine, he says, is that it appears to become less accurate the more qubits are added. “Classical computers simulating the machine show that when it gets to about 500 qubits the probability of getting the right answer approaches zero,” Morello says. In layman’s terms, no chance.
But Morello says we shouldn’t rush to write it off just yet. “I was a D-Wave sceptic and to a degree still am,” he says. “But having seen the capabilities, I would be careful of saying ‘that machine doesn’t work’. They could get things working.”
Others are more dismissive. “It’s not exactly science, what they’re doing. It’s high-level engineering, and I think it’s high-level salesmanship, too,” quantum information scientist Christopher Monroe of the University of Maryland, US, told National Public Radio just after the Google/NASA announcement.
D-Wave founder Geordie Rose is unconcerned by the criticism. “What we do is build computers,” he told NPR, “and if we can build the fastest computers the world has ever known, you can call them whatever you like.”