| （英） |
Decoding of surface codes can be done efficiently using minimum-weight perfect matching (MWPM). However, MWPM is not optimal for correlated X and Z errors, i.e, $Y$ error, such as depolarizing noise.
In this study, wemap the surface code decoding problem to energy minimization problem of the Ising model and propose a decoding method using simulated annealing to solve it. Through a series of numerical simulations, the proposed method achieves higher decoding accuracy than MWPM under depolarizing noise and shows comparable performance to decoding using CPLEX, which solves the combinatorial optimization problem exactly.Special-purpose hardware is now being developed to accelerate the approximate optimization of the Ising problem, and hence further speed-up is expected by using these hardware. |