| （英） |
Quantum state preparation is an essential subroutine in many quantum algorithms, such as quantum algorithm for Monte Carlo method, quantum machine learning.
In these algorithms, it is necessary to load classical data into the probability amplitudes of quantum states.
In recent years, several methods have been proposed for approximately loading data into quantum states.
One of these methods involves the use of tensor networks.
By using the Matrix Product State (MPS), which is one of the tensor networks, quantum states can be efficiently represented.
However, creating an exact quantum circuit corresponding to the MPS is difficult, and it is generally created approximately by limiting the bond dimension.
In this work, we aimed to reduce entanglement by performing a classical process of ``label swapping'' for the target probability distribution, to reduce the error from bond dimension limitations.
As a result, we found that the quantum circuit using MPS for the normal distribution and the log-normal distribution improves in approximation accuracy through ``label swapping''. |