Learning Mechanisms And Cryptographic Methods Based On Quantum Encoding Circuit Model

Abstract:Quantum computing,as a new computing method that have been theoretically proven the exponential acceleration effect compared with classical computing,has attracted much attention in the applications of var-ious fields,such as quantum cryptography,quantum chemistry,quantum biology and quantum finance,etc.Quantum computing is currently in the era of noisy intermediate-scale quantum(NISQ).Quantum circuits in NISQ can achieve quantum computing of about 50-100 qubits,and have proved quantum computing supremacy.However,the number of qubits in this era is limited and the noise is relatively large,which greatly hinder the realiza-tion of related algorithms and schemes of quantum computing.To solve the problems that the qubits are limited and the influence of noise is large in the current situation,this paper proposes the enhanced binary quantum encod-ing circuit model and the grouped coarse-grained boson sampling quantum encoding circuit model,and then designs quantum Siamese neural network and quantum hash algorithm based on the two encoding circuit models,re-spectively.We conduct simulation experiments and analysis on the quan-tum Siamese neural network and the quantum hash algorithm,which verify that the encoding circuit models can solve the current problems of the lim-ited number of qubits and the large noise impact.The main ideas of this paper are summarized as follows.(1)We propose the enhanced binary quantum encoding circuit model,which can use N+1 qubits to encode 2~Nbits of data.The enhanced binary quantum encoding circuit model exponentially reduces the use of qubits and provides the theoretical basis for the realization of quantum algorithms and applications on the promise of limited quantum resources.(2)We design a Siamese neural network based on the enhanced bi-nary quantum encoding circuit model and implement it on the Huawei HIQ quantum computing cloud platform.The experimental results prove that the proposed enhanced binary quantum encoding circuit model can use limited quantum bits to complete binary classification tasks.(3)We propose the grouped coarse-grained boson sampling quantum encoding circuit model,which can realize sampling grouping under the con-dition of multi-photon configuration,decrease the influence of boson sam-pling quantum circuit noise on the sampling results and reduce the experi-mental difficulty and sampling times of boson sampling.We design a hash function based on the grouped coarse-grained boson sampling quantum en-coding circuit model,which can successfully obtain a hash function that satisfies irreversible,Collision resistance and uniformity properties,and provide a new idea for cryptographic schemes based on quantum encod-ing circuit model.