Gaussian Boson Sampling with PennyLane

Gaussian Boson Sampling with PennyLane

This PennyLane demo introduces Gaussian Boson Sampling (GBS), a photonic sampling task that has been central to demonstrations of quantum computational advantage, including Xanadu's Borealis device and the photonic experiment of Zhong et al. It explains the physics behind boson sampling, from the bunching of bosons seen in the Hong-Ou-Mandel effect to Aaronson and Arkhipov's original proposal, and describes how the Gaussian variant, introduced by Hamilton et al., replaces hard-to-prepare single photons with experimentally friendlier squeezed states. Using the legacy PennyLane-Strawberry Fields plugin and its Gaussian backend, the demo builds a small circuit that squeezes four modes and passes them through a Haar-random four-mode interferometer of beamsplitters and phase shifters, then reads out the Fock-basis probability distribution. It verifies numerically that these probabilities match the theoretical formula, in which each event probability is proportional to the squared hafnian of a submatrix of the interferometer unitary, computed with The Walrus library. Because evaluating the hafnian is a sharp-P-hard problem, closely related to the matrix permanent, GBS is believed to be classically intractable at scale, which is exactly what makes it a compelling near-term demonstration of quantum advantage.

Run it

pip install -r requirements.txt
python demo.py

Source and license

Imported from demonstrations_v2/gbs/demo.py in PennyLaneAI/demos at c52c0abeb5122218aa96b38eea848864cce7323f, under the Apache License 2.0. Original authors: Xanadu and the PennyLane community. The upstream LICENSE is included alongside this example.