An Efficient General Decoding Algorithm for the Surface Code

摘要

A quantum error correcting protocol can be substantially improved by takinginto account features of the physical noise process. We present an efficientdecoder for the surface code which can account for general noise features,including coherences and correlations. We demonstrate that the decodersignificantly outperforms the conventional matching algorithm on a variety ofnoise models, including non-Pauli noise and spatially correlated noise. Thealgorithm is based on an approximate calculation of the logical channel using atensor-network description of the noisy state.