From phys.org >
“This works because the rules that govern how the ball will move, like gravity, are regular and known. But what if the rules changed randomly while the ball was on its way to you? In that case it’s next to impossible to predict the future behavior of that ball.
“And yet this situation is exactly what we had to deal with because the disintegration of quantum systems is random. Moreover, in the quantum realm observation erases quantumness, so our team needed to be able to guess how and when the system would randomly break.
“We effectively needed to swing at the randomly moving tennis ball while blindfolded.”
The team turned to machine learning for help in keeping their quantum systems – qubits realised in trapped atoms – from breaking.
What might look like random behavior actually contained enough information for a computer program to guess how the system would change in the future. It could then predict the future without direct observation, which would otherwise erase the system’s useful characteristics.
The predictions were remarkably accurate, allowing the team to use their guesses preemptively to compensate for the anticipated changes.
Doing this in real time allowed the team to prevent the disintegration of the quantum character, extending the useful lifetime of the qubits.
“We know that building real quantum technologies will require major advances in our ability to control and stabilise qubits – to make them useful in applications,” Professor Biercuk said.
Our techniques apply to any qubit, built in any technology, including the special superconducting circuits being used by major corporations.
“We’re excited to be developing new capabilities that turn quantum systems from novelties into useful technologies. The quantum future is looking better all the time,” Professor Biercuk said.