Exponential suppression of errors in Sycamore, a quantum processor designed by Google AI, is reported in Nature this week. This experimental demonstration may pave the way for the development of scalable, fault-tolerant quantum computers.
Quantum computers, like their classical counterparts, are prone to errors caused by ‘noise’ from the underlying physical system. A solution is to include, within the operation of the computer, a way to detect and correct the errors as they happen. One method of quantum error correction uses stabilizer codes, where multiple qubits (units of quantum information, equivalent to classical computer bits) are treated as a single logical qubit, allowing errors to be detected and corrected without damaging the information stored in the logical qubit. To achieve the potential of quantum computing, logical error rates need to be kept low.
Julian Kelly and colleagues investigate the performance of quantum error correction in the Sycamore processor, which consists of a two-dimensional array of 54 superconducting qubits. They run two stabilizer codes: a one-dimensional repetition code in a chain of up to 21 qubits that tests for error suppression, and a two-dimensional surface code of seven qubits as a proof of principle for setup compatibility with larger codes. The authors show that increasing the number of qubits that their repetition code is built on, from 5 to 21, leads to an exponential suppression of logical errors, up to 100-fold. This error suppression is stable over 50 rounds of error correction.
These results are encouraging because they suggest that quantum error correction can be successful in keeping errors under control. Although not yet at the threshold of error rates needed to realize the potential of quantum computing, the results here indicate that the Sycamore architecture may be close to achieving this threshold.
Sports: Little evidence that host countries win more Olympic medalsScientific Reports
Evolution: Group-living mammals may live the longestNature Communications
Education: Over one third of a year’s learning lost to COVID-19 pandemicNature Human Behaviour
Astronomy: Machine learning combs radio signals from spaceNature Astronomy
Animals: Cat-egorising play and genuine fighting in catsScientific Reports