Google Quantum Computer Works

    The search engine believes that the D-Wave machine has confirmed its principles of operation


    Quantum computers are computers that operate on the basis of quantum effects. Using the unusual properties of quantum superposition and quantum entanglement can give a significant leap in productivity. It can grow by several orders of magnitude. This growth can provide new opportunities in many areas of work with knowledge: artificial intelligence systems, simulation of various processes and forecasting.

    Therefore, many government research laboratories and IT giants are closely interested in this area. For example, IBM and Microsoftworking on quantum computing. Google also does not lag behind. In 2013, a search engine bought a quantum machine from D-Wave. This Canadian startup presents its product as the world's first commercially available quantum computer. The problem is that the scientists were not sure that the D-Wave chips really use quantum effects to do the calculations. December 8, Google announced that confirmation of this appeared.

    In 2013, Google bought a quantum machine from D-Wave. The computer was installed at NASA's Ames Research Center in Mountain View, California. For two years, researchers have studied the mechanisms of the quantum annealing methodand the possibility of their application. The operation of D-Wave devices was controversial among experts in quantum physics: they could not unanimously conclude that the devices really work as stated, and that they can achieve the promised computational speed.

    For example, among the skeptics is Matthias Troyer. A physicist from the Swiss Higher School of Zurich questions the speed of such machines. After researching the early D-Wave chips, his group came to the conclusion that the microcircuits did not provide any special advantages over traditional computers.

    Now Google claims they have evidence. The performance of a D-Wave machine and a conventional computer with a single processor was compared. A special task was created to prove significant superiority in computational abilities. A quantum computer coped with it 100 million times faster than usual. The graph shows a comparison of the performance of a conventional Monte Carlo (QMC) quantum computer and annealing simulation (SA) with a D-Wave quantum machine. On the right, when approaching 1000 binary variables, the advantage reaches 10 8 .




    However, even the head of the quantum laboratory, Hartmut Neven, admits that this task could be optimized in such a way that a regular computer would win or at least be at the same level. The quantum machine worked on the principle of quantum annealing , and a conventional computer used a simulated annealing . One could use what Neven himself calls a bug in the design of the D-Wave machine. Google says that the test shown is still important, because in the future ordinary computers will not have an advantage. It will disappear in the next generation of quantum machines.

    At Arxiv.orgGoogle published a scientific paper based on the results of this study. But so far no reviews have been written on it. Publications in scientific journals will follow. In the future, researchers want to improve some of the characteristics of quantum computers. For example, you need to increase the density and accuracy of the connections between qubits, as well as their connectivity. The qubits will teach to interact not only in pairs, but also in larger groups. Schematic representation of the connections between qubits in the chips of the D-Wave 2 family.




    The development of quantum computers can potentially make Google products better: there are tasks that are impractical to solve with conventional computers. But these are only distant prospects. It will take years to develop and implement such solutions. At the moment, D-Wave machines are extremely expensive, require large volumes of liquid nitrogen for cooling and are not universal. They can solve only a very specific set of tasks. What are these tasks, it still turns out.

    arXiv: 1512.02206 [quant-ph]

    Also popular now: