Is the world’s only commercial quantum computer really a quantum device, or a just regular computer in disguise? Controversy has long swirled around the computer produced by D-Wave, a company based near Vancouver, Canada. Now a paper published on the arXiv preprint server takes a step forward in showing that it really does operate on a quantum level.
D-Wave’s computer is a special type of quantum device: its quantum bits (or qubits) seek out a low-energy state that represents the answer to a given problem. Unlike a universal computer, this kind of computer, called an annealer, cannot answer any question thrown at it. Instead, it can only answer ‘discrete optimization’ problems. This is a type of problem where a set of criteria are all fighting to be simultaneously met, and there is one best solution that meets the most of them — one example being the simulation of protein folding, in which the system seeks a state of minimal free energy. The hope is that a quantum annealer should be able to solve these problems much more quickly than a classical one.
The company’s current top-line computer has 512 qubits. In some ways, this is miles ahead of work in universal quantum computers, where academics struggle to get just a handful of qubits to operate usefully. But even D-Wave admits that it doesn’t know exactly how its computer works, and critics have complained that it might not be quantum at all. Instead, it could be using classical physics to crunch calculations.
In 2011, a group led by scientists working with D-Wave published a paper in Nature with evidence that their 8-qubit system was working on a quantum level: it responded to temperature changes as expected for a quantum device. Now, a group of independent scientists follows that up by showing that the 128-qubit version of the D-Wave computer (or at least the 108 functioning qubits in the specific computer that they analysed) also seems to be behaving quantumly.
Simulations of quantum versus classical annealers show that a classical one has a fairly uniform probability of solving a problem correctly; a quantum device should instead have a low probability of success at solving hard problems, and a high probability of success solving easy ones. This is what they see with the D-Wave computer.
Scott Aaronson, a theoretical computer scientist at the Massachusetts Institute of Technology in Cambridge who has historically been sceptical of D-Wave’s claims, says that he is fairly convinced by the data, but that there are plenty of important questions remaining — including whether the current or future versions of the D-Wave computer will actually be any faster than classical machines.
The new paper, Aaronson notes, shows that a quantum annealer is actually expected to be slower than a classical one in many circumstances. “It may be that they really have built a quantum annealing device, which is academically very interesting, but that it provides no [speed] advantage. That may be the case,” says Aaronson.
The paper’s authors include several researchers from the University of Southern California in Los Angeles, which has a deal to use and experiment with the D-Wave computer recently purchased by aerospace company Lockheed Martin. The co-author contacted by this reporter declined to comment on the work until it appears in a peer-reviewed publication. As of March, that group now has a 512-qubit version of the D-Wave to play with, which could start to show a speed advantage over classical annealers.