In the experiment, the team used Jiuzhang to solve a problem that is challenging for classical computers. It used more than 200,000 samples to solve the problem.
The researchers, for the first time, used the quantum computer to implement and accelerate two algorithms – random search and simulated annealing – that are commonly used in the field of AI.
The fastest classical supercomputer in the world would take 700 seconds for each sample, meaning it would take nearly five years to process the same number of samples.
It took Jiuzhang less than a second.
In a synopsis article published by Physics, a magazine from the American Physical Society that reports on papers from the Physical Review journals, the editor wrote: “the result extends the list of tasks for which today’s noisy quantum computers offer an advantage over classical computers”.
“Previous claims of quantum advantage were challenged by suggestions that the quantum computer was not competing against the best-possible classical algorithm for the task,” the article said. “Whether the team’s quantum processor will still yield an advantage over classical algorithms optimised for solving graph problems is an open question.”
Related posts
Former China central bank official warns against miscategorising ‘virtual economy’
“My view is, we cannot simply categorise the service sector as the ‘virtual economy’,” Sheng said...-
China’s technology sector faces new round of lay-offs as global firms Ericsson, Tesla, Amazon and Intel continue to shed jobs
Swedish telecommunications equipment maker Ericsson’s plans in China will involve cutting about 240 positions at its... -
Taiwan’s incoming government faces tough task to win back public
The following week, mainland Chinese coastguard troops boarded and searched a Taiwanese cruise ship near Quemoy....
