ParityQC runs a 107-atom optimisation on QuEra's Aquila, helped by an error code

[StoneCold_HIAC]

A new experiment shows neutral atom quantum computers can correct some of their own mistakes while solving problems. Quantum Zeitgeist reports that ParityQC, the University of Innsbruck and QuEra ran an optimisation problem on QuEra's 107 atom Aquila processor

The team used extra atoms as a built in resource for error correction. Atoms were grouped into superatoms, each limited to a single excitation by the Rydberg blockade effect, forming a classical error correcting code they call the Rydberg Parity code

A new decoder called Cluster recovered target solutions with an average success rate of at least about 8 percent on the largest layout. Neither the raw data nor a classical greedy algorithm could reach that, and the gap widened as problems grew larger

The work is a preprint and has not yet been peer reviewed, and the 8 percent figure applies to one layout. It was published early on Saturday, just before the latest news window

Using the physics of the atoms themselves to protect against errors is an elegant idea, because it needs no extra hardware. An 8 percent success rate sounds low, but for hard optimisation problems that classical shortcuts miss, it can be valuable. Growing advantage with problem size is the most encouraging sign

Can analogue neutral atom machines solve useful optimisation problems soon? And how important is building error correction into the hardware itself?

ParityQC's QuEra Aquila QPU Runs Optimisation With 107 Atoms, Aided By Error Codequantumzeitgeist.com