New protocol is a huge breakthrough for the future of quantum computers

Friday, August 28, 2020

En français

Quantum computers will now have help tackling the central problem in their performance – noise.

Joel Wallman, a researcher at the Institute for Quantum Computing (IQC) and assistant professor of applied mathematics at the University of Waterloo, has developed a protocol that will help deal with the issue of noise in quantum computers so that they can tackle more complex problems.

“The intrinsic noise in quantum computers makes their output unreliable,” said Wallman, co-founder of Quantum Benchmark, a startup spun out of IQC. “So any problem that we know how to solve on a quantum computer can be solved better on conventional computers. To deliver quantum computers that can do something useful, we need to make larger quantum computers and work out how to accurately control them.”

Wallman, together with Robin Harper and Steve Flammia of the University of Sydney, has developed a new protocol that works on large systems – quantum computers running on many qubits (the quantum version of a classical computer’s binary bit)– that lets researchers characterize quantum noise across the qubits reliably and efficiently.

Prior to this work, researchers ran error assessment protocols that could only detect errors on a small subset of the qubits. The new method returns an estimate of the effective noise and can detect error correlations within arbitrary sets of qubits.

“The reason this protocol is so important is that if noise in systems don’t act locally, existing error correction and mitigation techniques just don’t work,” says Wallman. “And the data we obtained demonstrated that such nonlocal errors exist in real quantum computers.”

Wallman’s research team at the University of Waterloo and Quantum Benchmark is currently furthering the technique to characterize and suppress errors in specific data operations.

Efficient learning of quantum noise by Harper, Flammia, and Wallman was published in Nature Physics on August 10, 2020.

Tools based on the method are included in True-Q, the world-leading error characterization software from Quantum Benchmark.