Applications Of Classical Algorithms In Photonic Quantum Technology

In recent years,the cross-application research of classical algorithms and quantum information has received more and more attention.On the one hand,classical algorithms have provided many efficient and convenient research tools for quantum information processing,which deepens the understanding of quantum information science.On the other hand,the combination of quantum information and classical algorithms can catalyze more researches on "quantum-version algorithm".This thesis mainly focuses on convex optimization,machine learning and their applications in specific tasks of photonic quantum technology.The results include:1.Reconstruction of quantum channel via convex optimization.We reconstructed the seawater quantum channel via our convex-optimization-based quantum process tomography,showing that it could reveal the information of the channel more correctly.Next,we tested our method on the seven well-prepared quantum logic gates,which showed that our method surpassed the standard inversion and norm optimization both in robustness and correctness.In addition,we reconstructed a series of non-unitary channels,where our method was still able to reach up to 99.5% accuracy.Our method can reconstruct the quantum channel more robustly and precisely both in unitary and non-unitary situations,offering a more universal tool for further analyses on the study of quantum channel.2.Machine learning of Anderson localization.We demonstrated the Anderson localization on a 25×25 photonic lattice.We used convolutional neural network to classify the evolutional results of different length,in which we were able to identify the localized and non-localized states.In addition,we used support vector regression to predict the localization level,where the prediction results consisted well with the actual theoretical values.