This repository contains pre-generated data for the figure related to barren plateaus in variational quantum circuits (VQC), specifically focusing on the exponential and polynomial decay of gradients with respect to system size.

The original code used for the figure is available here. Generating the data using their provided Python scripts can be time-consuming, taking up to 2 days of continuous computation. To alleviate this burden, I have generated the data and uploaded it to this repository, along with minor fixes to ensure smooth execution.

data/: Contains the pre-generated data files.
barren_plateaus.py: Contains the original Python scripts with a minor fix.
figures/: Contains the figures generated from the script
README.md: This document.

The pre-generated data is located in the data/ directory. You can load and use this data directly in your analysis without the need to run the time-consuming scripts.

If you need to run the script to get data for different seed values for any reason, uncomment the get_data line in barren_plateaus.py and change the seed value. It will do the the whole computation again. Note the minor fix added:

import scienceplots

To run the scripts:

1. Ensure you have all dependencies installed: pip install -r requirements.txt

2. You need LaTeX to get the plots, for Linux install it using: sudo apt-get install dvipng texlive-latex-extra texlive-fonts-recommended cm-super

3. Execute the main script: python barren_plateaus.py

This project is licensed under the MIT License. See the LICENSE file for details.

Thanks to the original authors for their significant work on the classification of dynamical Lie algebras. This repository aims to build upon their foundation and provide easier access to their results.

Please cite the original authors if you use this data or any part of their code in your work:

@misc{wiersema2023classification,
      title={Classification of dynamical Lie algebras for translation-invariant 2-local spin systems in one dimension}, 
      author={Roeland Wiersema and Efekan Kökcü and Alexander F. Kemper and Bojko N. Bakalov},
      year={2023},
      eprint={2309.05690},
      archivePrefix={arXiv},
      primaryClass={quant-ph}
}