基于卷积神经网络的拓扑量子码译码研究

Decoding of Topological Quantum Codes Based on Convolutional Neural Network

由于量子纠错码的简并性以及传统匹配译码算法在错误相关性探测上的不足,拓扑量子码的译码问题一直以来都是很大的挑战。近年来,神经网络越来越多地应用到量子通信与量子信息中,极大地促进了量子科技的发展。这里,针对传统量子拓扑码的译码算法在错误相关性探测上的不足,搭建了卷积神经网络译码器,实现了拓扑量子码的高效译码。与传统匹配译码算法进行对比,提高了译码器对错误相关性的识别,实现了译码性能的提升。

Due to the degeneracy of quantum codes and the deficiency of traditional matching decoding algorithm with regard to the correlations of concerning errors,decoding of topological quantum codes has always been a great challenge.Recently,applications of neu⁃ral network in quantum communication and quantum information field are becoming popular.Here,in view of the shortcomings of tradi⁃tional decoding algorithms in error correlation detection,a convolutional neural network decoder is built.Compared with traditional matc⁃hing decoding algorithms,the ability of recognizing different errors of the decoder is clearly improved and a highly efficient decoding is achieved.