Research On Implementation Of Quantum Circuit Simulation Based On Open System

Quantum machine learning is a new method which combines machine learning algorithm with quantum computing technology.Its appearance marks the extension of classical machine learning algorithm to the quantum field,and becomes an important turning point in the development of artificial intelligence technology.As a common method to solve time series problems,hidden Markov model is one of the important models in traditional machine learning.However,like the traditional classical machine learning algorithms,the hidden Markov model also has the problems of insufficient computing resources and low efficiency in the training of large data and complex models.Therefore,making full use of the advantages of quantum computing,studying the implementation of quantum circuits based on quantum hidden Markov model can improve the computational efficiency and accuracy for some specific time series problems.At The same time,because the quantum hidden Markov model is naturally associated with the open system(non-unitary evolution)in quantum physics,so the study on the realization of quantum circuits of such non-unitary quantum evolution is the most popular noisy intermediate-scale quantum,NISQ)development to provide effective assistance.Based on the previous theoretical work of heuristic quantum hidden Markov algorithm,this paper adopts the classical+quantum hybrid quantum computing framework,proposes a set of quantum circuits and optimization methods to realize quantum hidden Markov algorithm,and carries out experimental verification on the time series data of different observed values,specifically including the following contents:·Based on the theoretical relationship between quantum hidden Markov model and the non-unitary time evolution of open quantum systems,a non-unitary evolution method is proposed to realize the non-unitary operation of quantum circuits,and is used to solve the Kraus operator of the parameters of quantum hidden Markov model.It lays a foundation for the study of quantum hidden Markov and realizes the transition from heuristic quantum hidden Markov model to fully quantized quantum hidden Markov model.·In the framework of hybrid quantum computing,quantum hidden Markov models are constructed for the time series problems of two observations and six observations respectively,and corresponding non-unitary quantum circuits and parameter optimization strategies are designed.·The time series data set is used to verify whether the quantum circuit model is reasonable,and two different parameter optimization strategies and training data amount are used to explore the factors that affect the model effect.The experimental results show that the quantum circuit model can effectively solve the quantum hidden Markov model parameters and realize the quantum hidden Markov process.The proposed quantum hidden Markov model under the hybrid quantum computing framework solves the problem of non-unitary Kraus operator for solving quantum hidden Markov model on quantum circuits.Compared with previous studies on quantum hidden Markov model,this model can simulate non-unitary time evolution on quantum circuits,which is helpful to the development of NISQ capability.