Chinese physicists challenge Google’s ‘quantum advantage’

A team in China claims to have made the first definitive demonstration of “quantum advantage” – by exploiting the counterintuitive workings of quantum mechanics to perform calculations that would be prohibitively slow on conventional computers.

They used beams of laser light to perform a calculation that had proven to be mathematically nearly impossible on normal computers. The team achieved in minutes what would take half the age of Earth on the best supercomputers in existence. Unlike Google’s first demonstration of quantum advantage last year, their version is virtually unassailable by any conventional computer. The results appeared in Science December 3¹.

“We have shown that we can use photons, the fundamental unit of light, to demonstrate quantum computing power far beyond its classical counterpart,” says Jian-Wei Pan of the University of Science and Technology of China in Hefei. He adds that the computation they performed – called the boson sampling problem – is not only a practical way to demonstrate quantum advantage, but has potential practical applications in graph theory, quantum chemistry and machine learning.

“It is certainly a tour de force experience and a milestone,” says physicist Ian Walmsley of Imperial College London.

The quantum advantage called into question

Teams from university labs and businesses have fought to demonstrate quantum advantage (a term that has now largely replaced the old “quantum supremacy”).

Last year, researchers at Google’s Quantum Computing Lab in Santa Barbara, California, announced the very first demonstration of quantum advantage. They used their state-of-the-art Sycamore device, which features 53 quantum bits (qubits) made from superconducting circuits kept at ultra-cold temperatures.².

But some quantum researchers have disputed this claim, on the grounds that a better classical algorithm that outperforms quantum might exist.³. And IBM researchers claimed that its classic supercomputers could in principle already run existing algorithms to perform the same calculations in 2.5 days.

To convincingly demonstrate the quantum advantage, it is unlikely that a much faster classical method could be found for the task under test.

The Hefei team, led by Pan and Chao-Yang Lu, chose a different problem for their demonstration, called boson sampling. It was designed in 2011 by two computer scientists, Scott Aaronson and Alex Arkhipov⁴, then at the Massachusetts Institute of Technology in Cambridge. This involves calculating the probability distribution of many bosons – a fundamental class of particle that includes photons – whose quantum waves interfere with each other in a way that essentially randomizes the position of the particles. The probability of detecting a boson at a given position can be calculated from an equation with many unknowns.

200 seconds

But the computation in this case is a “# P-hard problem”, which is even more difficult than notoriously tricky NP-hard problems, where the number of solutions increases exponentially with the number of variables. For several dozen bosons, Aaronson and Arkhipov have shown that there is no classic shorthand for the incredibly long computation.

A quantum computer, however, can bypass brute force computation by directly simulating the quantum process – allowing bosons to interfere and sample the resulting distribution. To do this, Pan and his colleagues chose to use photons as qubits. They performed the task on a quantum photonic computer operating at room temperature.

From the laser pulses, the researchers encoded information about the spatial position and polarization of particular photonic states – the orientation of the electromagnetic fields of photons. These states were then put together to interfere with each other and generate the distribution of photons that represents the output. The team used photodetectors capable of recording single photons to measure this distribution, which in fact encodes calculations that are so difficult to perform conventionally.

This way, Pan and his colleagues could find solutions to the problem of sampling bosons in 200 seconds. They estimate it would take 2.5 billion years to compute on China’s TaihuLight supercomputer – a quantum advantage of around 10¹⁴.

Practical issues

“This is the first time that a quantum advantage has been demonstrated using light or photonics,” says Christian Weedbrook, managing director of quantum computing startup Xanadu in Toronto, Canada, which seeks to build practical quantum computers based on photonics.

Walmsley says this quantum advantage claim is compelling. “Because [the experiment] being very close to the original Aaronson-Arkiphov scheme, it is unlikely that a better classical algorithm can be found, ”he says.

However, Weedbrook points out that for now, and unlike Google’s Sycamore, the Chinese team’s photonics circuit is not programmable, so at this point “it cannot be used to solve practical problems.”

But he adds that if the team is able to build a sufficiently efficient programmable chip, several important computational issues could be solved. Among these is the prediction of how proteins cling to each other and the vibration of molecules, says Lu.

Weedbrook notes that photonic quantum computing started later than other approaches, but now could “leap over the rest”. In any case, he adds, “It’s only a matter of time before quantum computers leave classical computers in the dust.”