We may unlock from the real true power of quantum computing is still far, but the D-Wave through its commitment to significantly upgrade the quantum processor let us taste a taste of the future.

When it is released early next year, the Canadian company’s new chip will be able to handle 2000 quantum quantum bits of data (qubits), almost the existing D-Wave 2X system twice the number of processors available, but will be better than the previous generation product processing speed 1000 times faster.

D-Wave machine are millions of dollars worth of computers, which use “quantum transistors” to process the data, in addition to the use of liquid nitrogen to cool the tiny ring of niobium to absolute zero. Only a few places in the use of this system, including Google and Universities Space Research Association (Universities Space Research Association), Lockheed Martin (Lockheed Martin) and Los Alamos National Laboratory (Los Alamos National Laboratory). In addition, D-Wave also provides Cloud services allow users to access its quantum computer system. 38634-5108

Most of quantum computing is still in the theoretical stage, its main research how strange and counterintuitive way to develop a very powerful machine at the atomic level. For certain tasks, a quantum computer has the potential to be significantly more than the existing system is many times faster, and more energy efficient. But GM does not currently exist in the quantum computer, D-Wave systems with different atomic behavioral characteristics (such as entanglement and superposition state) to solve a series of difficult computational problems.

“We chose the direction of development is very rapid, almost every year the number of quantum bits will double.” D-Wave business development and strategic partnerships Colin Williams said.

Lift D-Wave quantum processor makes the system a bit closer to have the ability to challenge traditional computers, while new processor also supports adding new features (allows for more efficient operation). tda7580

“From internal testing, this is a very worthwhile thing. By using this feature we have accelerated 1000 times to solve those problems.” Cambridge CW TEC’s William says.

If you follow today’s computer analogy, then, D-Wave system is not a general-purpose computer.

D-Wave can do far more than any of you have requested a computing tasks, it is designed to deal with specific tasks (like unconstrained binary Optimized sampling and related issues). Such optimization issues can give a very simple example, you have to develop a plan of the house, it is possible to meet your ideal of conditions, details, but while keeping in your budget range.

D-Wave specific work a processor can execute can also be applied to a number of areas, particularly in the training of machine learning models.

However, the establishment of universal quantum computer enormous difficulties remain, there are several unsolved engineering problems before. UCL nanoelectronics and photonics chip Professor John Morton by past trends projections, forecast first-generation universal quantum computing processor will not appear until the 2030s.

John Morton mention just a calculator and words can not be called a computer, so the D-Wave systems can not be said to be a universal quantum computer. When “A calculator can solve many specific problems. A lot of people are using it, and you can use it in various industrial fields. Therefore, when D-Wave demonstrated a number of industrial applications can use their machines perhaps indeed it can be applied to the field, but it is still only a specific device only. ”

Although Google does not currently apply to the D-Wave machine learning system itself, but the credibility of D-Wave Processor (academics have been questioned) verified in a test last year, Google’s.

In Google’s experiment, when dealing with the same size as the amount of data required for D-Wave annealing time is several orders of magnitude less, the following figure shows the simulated annealing (SA), quantum Monte Carlo method and the D-Wave 2X in execution annealing operation time by comparing the results.

The experiment found that D-Wave 2X processor performing an operation similar to the conventional processor faster than 1 billion times, but Williams said that more importantly, it proves the feasibility of D-Wave chips in the future.

Experiments also show us the D-Wave 2X processor handling weak – strong cluster layout aspects of network problems several instances (see figure below). It is shown in three different size of network 295,490,945 qubits. Each cluster consists of an 8-bit quantum unit Chimera map. Orange dot is described by a strong local field of quantum bits, quantum bits blue-green dots represent the weak field. Blue lines represent magnetic coupling, the red line represents the strong anti-magnetic coupling. Note that since not all of the 1152 qubits are operational, the resulting image may appear irregular.

“The main results of this experiment is not entirely the speed increase, because in other classical algorithm can do better. This experiment shows that quantum tunneling is really happening in the D-Wave chip. This shows that even if the range of tunneling limited, it is still a useful calculation tool. ”

“Google, like us understand that when we make further chip more closely connected, the current well-run classical algorithms will soon complete failure.” Williams also mentioned.

D-Wave has been there from many investors to raise millions of dollars, investors including Goldman Sachs investment bank, In-Q-Tel (CIA Investment), Bezos Expeditions (Amazon founder Jeff Bezos) and BDC Capital , Harris & Harris Group and DFJ.

A machine capable of imitating human speech

D-Wave chip too specialized, but to restrict its usefulness, which is another critical perspective on the chip, however, Willams not agree with this view.

“That D-Wave single chip uses too, I do not agree with this view. In fact, this kind of chip in one area can be extended to more different areas,” he said.

Williams did not directly mention which companies in this way using the D-Wave chip, he said, this type of processor is used in business transactions to optimize the track, in the biological sciences to study how proteins are combined and build protein sequence filter to accurately detect every potential protein sequences, D-Wave chip, this application helps to improve the detection of terrorists lists security services, while also helping to develop AI and computer vision binary classifier.

However, the theoretical basis of D-Wave chip unsupervised machine learning, training data to provide a neural network, pattern recognition by machine learning. Williams believes, D-Wave processor will have a huge impact, Google may be able to explain the research interest in the art.

“We believe that this is the machine learning and AI machine optimal application case .D-Wave chip will make it possible for the field of machine learning, especially for the generation of unsupervised learning research, bring about a complete revolution,” he says .

“With this quantum chip, we will have the ability to solve challenging field of machine learning problems appearing earlier – ‘How to make unsupervised-forming machine learning can be effectively put into practice?’”

If we can solve these problems, you will be able to apply machine learning to achieve many incredible breakthrough. By training machine, you can make it generate new data from statistical terms, the new generation of data can be achieved with the data used to train the machine itself highly similar effect.

Williams predicted that the future of D-Wave chip can train a machine, for example by painting masters to train the machine, the machine is capable of generating new works have artistic value, or enable the machine to highly realistic imitation of the human voice.

D-Wave chip has been used in the machine learning experiments conducted to construct a kind of Boltzmann machine, a stochastic recurrent neural network can also be called “Quantum Boltzmann machine”, Willams said, “This species quantum Boltzmann machine is different from the previous machine learning models “in nature.

Williams believes, D-Wave chip, or other types of quantum processor will not replace some of the classics of computer chips, on the contrary, will be used together with these classic chip users.

“We believe that quantum computing will not replace the classic machines, this calculation method to change the world by strengthening the classic system”, Williams expressed.

For example, you can get a quantum computer’s output and will be used as input to a heuristic search algorithm. The key point is that quantum algorithms can help you get closer to not be able to find a good solution, a classic algorithm is able to contribute to ultimately find such a solution.

We can also choose a difficult problem from a number of pre-processing technology, the use of such a quantum chip this problem into a series of small problems.

Williams said that in addition to handling qubits 2000, D-Wave will be based on all the experience we have learned, the use of a new type of topology design a “next-generation chips.”

For more detailed information about the new D-Wave chip

For those professionals interested in the substance of the issue, Williams on 2,000 properties qubit chip made depth explanation.

Each designed to a D-Wave quantum annealing processor is used, the use of quantum physics knowledge to find a state of minimum energy consumption, which will help to solve optimization problems and related sample problem mentioned above. The new chip on how to more effectively control the annealing process, Williams also made a detailed explanation.

After reference Hamiltonian equation (in the state of a given system conditions, the use of this formula can output energy in the system), he said, “This new chip is not only capable of handling more qubits, we have to use it changed a lot of other features. using the previous D-Wave chip, we can only observe one kind of track annealing. currently, we can only do basic initial Hamiltonian closed loop, open loop Hamiltonian final. ”

Now use qubits capable of processing 2,000 chips, we will be able to better control parameters to control the trajectory.

We can use a lot of features abort the annealing process, and be able to quickly accelerate the process. It is no longer required at a constant rate of an annealing process.

It will be very interesting, because you can progress in the annealing process, the detection of quantum state, which is a quantum Boltzmann machine is a very important feature.

We also have a rapid thermal annealing previous generation system, it is possible to complete the operation within 20 microseconds annealing process. However, the use of our new system can be achieved in 5 microseconds to complete this operation.