sakkmesterke -

Why quantum needs a classic approach for supremacy

Google claims it has developed an algorithm for a quantum computer that would take a traditional “classical” computer 10,000 years to run. We investigate

There has been a flurry of activity in the last week of September. First, IBM unveiled a new facility in New York which will support a 53Qubit (quantum bit) system. Then news broke in the FT over the weekend about a new scientific paper from Google discussing how researchers at the internet search giant have achieved quantum supremacy – a term used to describe the situation where an algorithm would be almost impossible to run on a classical computer architecture compared with using a quantum computer.

According to the FT, the paper describes how a Google quantum computer “was able to perform a calculation in three minutes and 20 seconds that would take today’s most advanced classical computer, known as Summit, approximately 10,000 years”.

This week, D-Wave Systems also announced plans to make its Advantage quantum computer available to business users via the cloud. According to D-Wave, the Advantage quantum system will power a new hardware and software platform that will accelerate and ease the delivery of quantum computing applications.

“Quantum computing is only as valuable as the applications customers can run,” said Alan Baratz, chief product officer at D-Wave. “With the Advantage quantum system, we are building the first-ever quantum computer designed to deliver business benefit.

“Our investments across our quantum platform, which includes the Leap quantum cloud service, the Advantage quantum system and the Ocean developer tools, will together allow customers to solve even more complex problems at greater scale and bring emerging quantum and hybrid applications to life.

“Our ongoing efforts to further productise and commercialise our quantum platform are good for the growing ecosystem, good for the quantum computing market and, most importantly, good for our customers, who are building the first commercial quantum applications.” 

Supreme architectures

IBM believes quantum computers will never reign “supreme” over classical computers, but will rather work in concert with them, because each have their unique strengths. The idea is to run algorithms on a hybrid architecture.

Dario Gil, director of IBM Research, said: “For quantum to positively impact society, the task ahead is to continue to build and make widely accessible truly programmable quantum computing systems that can implement, reproducibly and reliably, a broad array of quantum algorithms and programs. This is the only path forward for practical solutions to be realised in quantum computers.

“Only then will we get to the era of quantum advantage, when we will get to use ‘quantum + classical’ systems in concert to accelerate discovery in science and to create commercial value in business. That goal, no matter current claims, has not been realised.”

Generally speaking, a quantum computer does not offer the precision of a classical computer architecture, which relies on binary, 0 and 1, yes and no decisions.

Stefan Woerner, global leader for quantum finance and optimisation at IBM, said: “Classical computers use binary optimisation and make many yes/no decisions that have to be correlated. Whenever you add a binary variable to the problem, you double the number of checks.”

In practice, this means that when attempting to solve a problem that has several variables, the computations needed to run these correlations grows exponentially. However, Woerner added: “Some problems can be formulated in a way similar to quantum chemistry.”

This is the domain of quantum computing and, for companies like IBM, it can be applied in areas such as quantum mechanics, genomics, supply chain optimisation and financial risk models.

According to Woerner, Monte Carlo simulations for finance is an example of an algorithm where there is a proven advantage in using a quantum computer over a classical approach. The idea is to randomly generate future price developments based on statistical models. By running the simulations millions of times, it is possible to evaluate price scenarios, said Woerner, adding: “You can estimate a risk matrix, but to improve the estimate, you have to double the number of samples of data.”

Read more about quantum computing

  • Volkswagen Group recently demonstrated how it was working with D-Wave to research uses for quantum computing. We find out what the car maker is hoping to achieve.
  • Industry experts predict it will take 10 years for quantum computing to become a reality, but Microsoft believes it has the research edge, with systems, software and technology to get there in five.

Some of these calculations can run for an extremely long time. In Woerner’s experience, a large portfolio of loans to assess credit risk can take hours to days to run on a classical computer architecture. “We developed a quantum algorithm that can speed this up,” he said.

According to Woerner, the quantum computer algorithm may require only a few thousand samples, compared with the millions that need to be processed in a classical computer. “Quantum interference can cancel out wrong answers and amplify the answers we are interested in,” he said. “This allows us to run some algorithms more efficiently.

“We estimate it would take 15 minutes to run a Monte Carlo simulation for a portfolio of a million assets. But if we optimise, it could be near real time.”

But there is a fundamental problem holding back deployment of such systems to take advantage of quantum computing, said Woerner. “Quantum computers today are noisy,” he pointed out. “There are errors in their operations and you can accumulate too many errors.”

The systems for quantum computing need to be designed in a way that can cope with these errors, he said. “In the long run, we will use error correction using physical Qubits. But the goal of total error correction is many years into the future. Today we do the best on noisy machines.”

IBM sees hybrid quantum algorithms as an approach that can be used today to provide a level of error correction. Generally speaking, such a system combines a feedback loop between a classical computer and a quantum computer to improve the accuracy of the algorithm being executed.

Read more on Chips and processor hardware

Data Center
Data Management