A red laser illuminates components of a quantum boson sampling machine. (Courtesy James Gates / Science )
While studies into the bizarre world of quantum mechanics have raised the promise of staggeringly powerful computers, scientists are finding it extremely difficult to build a device that comes even close to outperforming conventional computers.
Recently, however, researchers said they had built a comparatively simple machine that could very soon run circles around classical computers -- at least when it comes to solving one single quantum puzzle.
The so-called quantum boson sampling machine is based on a phenomenon known as the random quantum walk -- the mysterious path that light particles take when traveling along a series of paths and obstacles. In three studies published this week in the journal Science, scientists proposed that the device and other quantum walk experiments were potential precursors to a quantum computer.
"One reason that these experiments are important is that this approach may be the first to solve a specific problem faster than a classical computer," said James Franson, a physics professor at the University of Maryland Baltimore. Franson was not involved in any of the three studies, but wrote a Perspectives piece on their significance in the same issue.
What makes the world of quantum physics so unlike daily reality is that extremely small particles can exist in two places at once, among other seemingly contradictory qualities. It is this uniqueness that leads scientists to theorize that a computer based on quantum particles would be capable of solving multiple tasks simultaneously.
A photon, a single packet of light, is a type of boson. The boson sampling machine sends photons into a network of channels, where they encounter mirrored beam splitters that divert the photon's path. A series of photon detectors document just where the photon exits the machine, as it did not follow a straight line.
The machine is roughly analagous to a Galton board -- an experimental device that resembles an old arcade game in which a coin is dropped into the top of a box filled with metal pins. The coin bounces off each pin as it nears the bottom of the box, and falls into one of a number of small slots.
While the path of the coin is erratic as it bounces from pin to pin, most of the coins will land in the middle slots, and fewer will land in the corner slots. If you drop enough coins, the distribution will resemble a bell curve as they collect in the bottom of the board.
However, in the quantum world, a particle does not merely fall to the left or right of a pin -- it can fall on BOTH sides of each pin it strikes. Because of this, the particles distribute themselves very differently at the end of their journey.
It turns out that today's computers have a difficult time calculating the probable pattern of distribution of quantum particles if a lot of them are sent through the machine. It takes longer and longer to solve the problem as each photon is added.
Operators of the boson sampling machine, though, can arrive at the solution much more quickly. They simply need to examine the photon detectors to determine the final placement of the photons and then describe their arrangements in terms of probability.
Researchers have yet to build a sampler large enough to prove their hypothesis. (Current devices use a handful of photons, whereas a large number would be considered 20 to 30 photons.)
Nevertheless, they said that day was not far off.
"Scaling this to large numbers of photons will be a much simpler task than building a universal quantum computer," wrote Matthew Broome, the lead author of one study and a quantum information researcher at the University of Queensland in Brisbane, Australia.
Return to the Science Now blog.
Follow me on Twitter @montemorin