Quantum computers have the potential to offer ground-breaking computational power, but only if protected from noise. This is done via error correction; a way to exchange many noisy qubits (units of computation) for fewer but more perfect qubits.
The crucial subtask of monitoring measurements from the quantum processor to deduce when errors have occurred is called 'decoding'. This must be performed extremely quickly in order to keep pace with the quantum machine.
In the paper 'Actis: a strictly local union-find decoder', Tim Chan and Professor Simon Benjamin modified an existing decoding algorithm to make it local (ie runnable on a grid of identical processing cells, each communicating with their nearest neighbours only) - locality has various practical benefits in speed, layout and robustness. The authors tested their local design and found that its runtime behaved more favourably than the original algorithm, and suggest the use of 'asynchronous' hardware to maximise their design's absolute performance.