Quantum Computing with Photonic Boson Sampling

Photonic boson sampling is a specialised approach to quantum computing that exploits the statistical complexity of indistinguishable photons traversing a linear optical network. By injecting single photons or squeezed states into a programmable interferometer, the device samples from probability distributions that are believed to be intractable to simulate classically. This model relies on multi-photon quantum interference and the inherent bosonic nature of photons to generate output patterns corresponding to matrix permanents or Hafnians, which encode computationally hard problems. Although not universal in the gate-based sense, boson sampling and its Gaussian variant exemplify an early path to demonstrating quantum advantage and exploring applications ranging from graph theory to molecular modelling. Advances in photon-source efficiency, interferometer stability and detection have propelled boson sampling from small-scale demonstrations to larger-mode devices capable of registering tens or hundreds of photons. This progression highlights the potential of photonic platforms for both near-term quantum experiments and the benchmarking of quantum computational supremacy.

Research from Nature Portfolio

Recent developments have produced a cloud-accessible photonic processor integrating high-efficiency quantum-dot single-photon sources with a reconfigurable linear-optical circuit and machine-learned error mitigation. This platform accommodates both gate-based logic and native photonic operations, achieving six-photon boson sampling and demonstrating quantum neural network classifiers on the same chip. A time-bin-encoded Gaussian boson sampler with adjustable squeezing parameters and a programmable interferometer has been used to solve graph-based problems in drug discovery, doubling the success probability of clique-finding relative to classical sampling and enabling molecular docking and RNA-folding prediction. Foundational work has also validated quantum advantage with a dynamic, time-multiplexed processor executing Gaussian boson sampling on over 200 modes, registering hundreds of photons and surpassing classical runtimes by orders of magnitude, thereby affirming the scalability of photonic sampling experiments.

Quantum Computing with Photonic Boson Sampling publication trend

The graph below shows the total number of articles in quantum computing with photonic boson sampling across all publications each year (not limited to Nature Index journals).

Technical terms

Boson sampling: A non-universal quantum protocol in which indistinguishable photons traverse a linear interferometer and yield output patterns related to the permanent of a matrix, believed to be hard to simulate classically.

Gaussian boson sampling: A boson sampling variant utilising squeezed vacuum states and measuring photon-number statistics, linked to Hafnian matrix functions and employed in graph-theoretic and molecular applications.

Linear optical network: An arrangement of beam splitters and phase shifters that implements a unitary transformation on input optical modes without direct photon–photon interactions.

Quantum advantage: The demonstration that a quantum device can solve or sample from certain problems more efficiently than classical computers.

Photonic quantum interference: The indistinguishable overlap of photon probability amplitudes leading to non-classical correlations in detection events.

References

1. A versatile single-photon-based quantum computing platform. Nature Photonics (2024).
2. A universal programmable Gaussian boson sampler for drug discovery. Nature Computational Science (2023).
3. Quantum computational advantage with a programmable photonic processor. Nature (2022).
4. Quantum-Inspired Classical Algorithm for Graph Problems by Gaussian Boson Sampling. PRX Quantum (2024).
5. Molecular docking with Gaussian Boson Sampling. Science Advances (2020).
6. Quantum sampling problems, BosonSampling and quantum supremacy. npj Quantum Information (2017).