photon-catalysis

Code for "Photon catalysis for general multimode multi-photon quantum state preparation"

Andrei Aralov, Émilie Gillet, Viet Nguyen, Andrea Cosentino, Mattia Walschaers, and Massimo Frigerio arXiv:2507.19397

---

Installation

Simply run this line to install the package using PIP:

pip install -e git+https://github.com/EQ15T/photon-catalysis.git#egg=photon_catalysis

For development:

git clone https://github.com/EQ15T/photon-catalysis.git
cd photon-catalysis
pip install -e .

Optionally, to run the examples simulating the boson sampling implementation of our schemes, the Perceval library from Quandela needs to be installed, using the following command:

pip install -e git+https://github.com/EQ15T/photon-catalysis.git#egg=photon_catalysis[boson_sampling]

Or:

pip install -e .[boson_sampling]

Similarly, the scripts simulating the gaussian boson sampling implementation requires the StrawberryFields library from Xanadu:

pip install -e git+https://github.com/EQ15T/photon-catalysis.git#egg=photon_catalysis[gaussian_boson_sampling]

Or:

pip install -e .[gaussian_boson_sampling]

Note that both can be simultaneously installed with the [boson_sampling,gaussian_boson_sampling] argument.

---

Examples

Basic usage

Using this software, one could obtain the set of linear forms in creation operators, such that their product, upon conditioning, corresponds to the desired core state. For example, for the state

$$\ket{\psi} \propto \ket{2000} + \ket{0200} + \ket{0020} + \ket{0002}$$

the following code could be used:

from photon_catalysis.utils import kets_to_state_dict
from photon_catalysis.optimal_preparation import optimal_preparation

extra_photons = 2
state = kets_to_state_dict([(2, 0, 0, 0), (0, 2, 0, 0), (0, 0, 2, 0), (0, 0, 0, 2)])
for W, prob, fid in optimal_preparation(state, extra_photons, num_decompositions=10):
    print(W)

Rows of W define the linear forms, extra_photons is the number of catalysis photons. test_photon_catalysis contains tests that could serve as more complete usage examples.

Boson sampling implementation (DV)

A Perceval circuit (and the accompanying input state and post-selection rules) can then be generated, and simulated, with:

circuit = StatePreparationCircuit(W, state)
pcvl_circuit, input_state, post_select = circuit.to_perceval(
    photon_addition_r=0.7, decompose_unitaries=False
)
simulation = pcvl.Simulator(pcvl.SLOSBackend())
simulation.set_circuit(pcvl_circuit)
simulation.set_postselection(post_select)
final_state = simulation.evolve(input_state)
print(final_state)

---

Additional scripts

The photon_addition_performance_bs.py script from the scripts directory illustrates, on selected states, the trade-off between fidelity and probability of success resulting from the non-ideal implementation of photon addition with a beam-splitter, in the (DV) boson sampling setting. The script can be run from the command line, and outputs data and plots to the results directory.

The photon_addition_performance_sqz.py script is its counterpart for the Gaussian boson sampling scheme, simulating the imperfect photon addition achieved by two-mode squeezing and PNR detection. Note that this simulation is slow and the script can take hours to run and complete.

The probability_plot.py script produces the data needed to reproduce figure 3 in the manuscript.

The paper_tables.py script reproduces the main result table (table IV).