Google's latest quantum device achieves its goal of error correction

The latest generation of quantum chips from Google Quantum AI have been shown to exhibit quantum error correction that suppresses faults below a critical threshold required to achieve future practical quantum computing applications, according to an article published this week in Nature. Scaling the device performance could help meet the operational needs of fault-tolerant quantum computing on a broad scale.

In some tasks, quantum computing may be able to outperform traditional computers and speed up computation. Current prototypes of quantum computers, however, cannot operate for long enough to produce useful results due to their error-prone nature.

In order to overcome this, researchers in quantum computing have developed a method known as quantum error correction, in which errors can be found and corrected without compromising computation by distributing information among numerous qubits, which are units of quantum information that resemble classical computer bits. Quantum error correction may introduce more faults than it fixes due to the qubit overhead it requires.

Because of this, it has been difficult to achieve "below threshold" operations, in which the uncorrected error rate is below a key amount necessary for error correction to function as intended and exponentially decrease errors.

The latest iteration of a superconducting quantum processing chip architecture, Willow, is presented by Hartmut Neven and associates. It can perform quantum error correction below the critical threshold for a specific quantum error correcting technique called surface code. Their technology maintains its functionality and decodes mistakes in real time while operating for up to 1 million cycles over several hours.

A 72-qubit processor and a 105-qubit processor are used by the authors to run surface codes. The logical error rate is cut in half each time the coding distance is extended from 3 to 5 to 7. The authors argue that the basis for executing large-scale quantum algorithms with error correction is this exponential reduction of logical mistakes.