QC-DFT

This repository provides source code used to generate the data presented in "Efficient Classical Computation of Single-Qubit Marginal Measurement Probabilities to Simulate Certain Classes of Quantum Algorithms" paper.

How to use

Setup

• Clone the repository with the command:

git clone https://github.com/BRIN-Q/QCDFT-ML.git

• Move into the repository directory

cd QCDFT-ML

• Build the virtual environment and directories using BUILD.py. Ensure that you have at least Python 3.11 with pip and venv installed.

python BUILD.py

Generating Figure 1 & S1

To generate Fig. 1 & Fig. S1 in the paper, we need to do the following steps:

• Generating neural network training data

  To do this, we simply need to run all of the cells in Generate-Training-Data.ipynb

• Creating and training the neural networks

  To do this, we simply need to run all of the cells in Train.ipynb. The resulting training data will be stored in the Results folder and can be plotted using Generate-Figures.ipynb to generate Fig. S1.

• Test the SQP error, mean fidelity, and mean SQP

  To do this, we simply need to run all of the cells in SQP.ipynb. The resulting data will be stored in the Results folder and can be plotted using Generate-Figures.ipynb to generate Fig. 1.

(Update) Generating Figure 1 with Tensor Network Simulation

• Training the neural networks

  Generate training data and train the neural networks using Generate-Training-Data.ipynb and Train.ipynb as in the previous section.

• Test the SQP error, mean fidelity, and mean SQP of the tensor network simulation using various maximum bond dimension.

  To do this, set the maximum bond dimension in Generates-Figure1-Data.ipynb then run the whole notebook. We can also choose the contraction method provided by the pennylane's default.tensor device ("auto-mps", "swap+split", or "nonlocal"). Additionally, Generates-Figure1-Data.ipynb calculates SQP error, mean fidelity, and mean SQP of the Bernardi and our proposed 1-RDM method.

• Generating Figure 1

  To do this, run all of the cells in Generate-Figure1.ipynb. We can select the simulation that we want to plot in the last cell. The available methods for the last cell are "1-RDM" and "MPS". If we choose "MPS", then we need to specify the contraction method and the maximum bond dimension as used in the previous step. While if we choose "1-RDM", simply add blank input "_" for last two function arguments.

Generating Table 1 & Table 2

To generate Table 1 & Table 2 in the paper, we simply need to run all of the cells in Grover.ipynb.

Generating Figure 2

To generate Fig. 2 in the paper, we need to run Shor.py. To do this, first, we need to edit the python file to specify the number of threads to use by modifying the line below to the desired number of threads

thread_count = 16

Then, we should activate the virtual environment and run the python file by using these following commands (on Linux)

source qc-dft-venv/bin/activate
python Shor.py

The resulting data will be stored in the Results folder and can be plotted using Generate-Figures.ipynb to generate Fig. 2.

Generating Figure S2

To generate Fig. S2 in the paper, we simply need to run all of the cells in Time.ipynb. The resulting data will be stored in the Results folder and can be plotted using Generate-Figures.ipynb to generate Fig. S2.