On the learning abilities of photonic continuous-variable Born machines

This repository contains the source code for the article "On the learning abilities of photonic continuous-variable Born machines".

The simulations are executed with Python 3.10.13, and the requirements are found in requirements.txt.

For executing the scripts, please install requirements with pip install -r requirements.txt. to ensure that the correct version of piquasso is used, it is advised to use a virtualenv. Note also, that matplotlib uses TeX for the plots.

The data for the 2-mode simulations can be obtained by

python3 d_2_script.py

which saves the data in the data/ folder. The figures in the article can be generated by

python3 d_2_plot.py data/<FILENAME>

Similarly for the 3- and 4-mode simulations.

The gradient estimation script from Fig. 6 can be executed with

python3 gradient_estimation.py

For the benchmark in Fig. 1, execute

python3 benchmark.py

which saves the data in the data/ folder, and execute it with

python3 benchmark_plot.py data/<FILENAME>