---
slug: 2026-arxiv-1912.07634-applications-near-term-photonic-quantum-computers
title: 'Applications of near-term photonic quantum computers: software and algorithms'
source: arXiv
link: https://arxiv.org/abs/1912.07634
category: app
triage_date: '2026-06-07'
triager: port_until_julien_parallel
bootstrap_date: '2026-06-07'
---

# Dossier: Applications of near-term photonic quantum computers: software and algorithms

| Field        | Value |
|--------------|-------|
| Slug         | `2026-arxiv-1912.07634-applications-near-term-photonic-quantum-computers` |
| Source       | arXiv |
| Link         | [https://arxiv.org/abs/1912.07634](https://arxiv.org/abs/1912.07634) |
| Category     | app |
| Triaged      | 2026-06-07 by port_until_julien_parallel |
| Bootstrapped | 2026-06-07 |

## Sourcing (Romain)

### Domain classification

**Primary: `optimisation`** (graph-based combinatorial problems). The bulk of the
paper develops Gaussian Boson Sampling (GBS) algorithms for graph optimisation:
dense subgraph identification and maximum clique are the canonical, most
distinctive use cases, both encoded via the adjacency matrix to bias the
sampler toward dense structures.

Secondary domains, both materially present:

- **`quantum chemistry`**: the vibronic-spectra section maps the GBS device
  onto the Franck-Condon problem (molecular vibronic transition profiles),
  the original physics motivation for boson sampling with squeezing and
  displacement.
- **`machine learning`**: graph similarity produces GBS feature vectors / graph
  kernels for graph classification, and the point-processes section is a
  sampling/statistics primitive with ML uses.

This is a software-and-review paper (the Strawberry Fields `apps` layer), so it
deliberately spans domains; the unifying object is the GBS sampling primitive
rather than any single application.

### Expert persona for Julien

A quantum-optics / photonic-quantum-computing specialist who works at the level
of the GBS device model, not just the application maths. They should read the
squeezing / linear-interferometer / displacement decomposition fluently and map
it to concrete hardware demands: number of optical modes $m$, per-mode squeezing
parameters $r_i$ (via $\tanh r_i = c\lambda_i$ from the Takagi-Autonne
decomposition), target mean photon number $\bar{n}$, and the detector model
(photon-number-resolving versus threshold "click" detectors, which select
between the hafnian, loop-hafnian, and Torontonian output laws). Calibrated
intuition about Xanadu's Strawberry Fields / The Walrus stack and about
real-device limits (squeezed-light source quality, photon loss $L$, mode count)
is what lets them turn this review's abstract device into a portability
fingerprint. Comfort with the Franck-Condon / vibronic-spectra mapping is a
useful secondary skill for the chemistry section.

### Related prior work (brief bibliography)

- Hamilton et al., *Gaussian Boson Sampling*, PRL 119, 170501 (2017).
  [arXiv:1612.01199](https://arxiv.org/abs/1612.01199). Defines the sampling
  primitive this whole paper builds on (the hafnian output law).
- Huh et al., *Boson sampling for molecular vibronic spectra*, Nature Photonics
  9, 615 (2015). [arXiv:1412.8427](https://arxiv.org/abs/1412.8427). The
  chemistry application underlying the vibronic-spectra section.
- Arrazola & Bromley, *Using Gaussian Boson Sampling to Find Dense Subgraphs*,
  PRL 121, 030503 (2018). [arXiv:1803.10730](https://arxiv.org/abs/1803.10730).
  Source algorithm for the dense-subgraph / optimisation sections.
- Schuld et al., *Measuring the similarity of graphs with a Gaussian boson
  sampler*, Phys. Rev. A 101, 032314 (2020).
  [arXiv:1810.10644](https://arxiv.org/abs/1810.10644). The graph-kernel /
  ML application formalised.
- Killoran et al., *Strawberry Fields: A Software Platform for Photonic Quantum
  Computing*, Quantum 3, 129 (2019).
  [arXiv:1804.03159](https://arxiv.org/abs/1804.03159). The underlying library
  that this `apps` layer sits on; needed to read the software claims.
- Zhong et al., *Quantum computational advantage using photons* (Jiuzhang),
  Science 370, 1460 (2020).
  [arXiv:2012.01625](https://arxiv.org/abs/2012.01625). Forward pointer: the
  same GBS primitive realised on real photonic hardware, the natural
  portability comparator to this software-era review.

### Difficulty estimate

**medium.** The paper is open-access on arXiv and the full LaTeX source plus
figures are already in `paper-source/`, so there is no paywall or SI-retrieval
blocker. It is self-contained and pedagogical: the GBS device model, the
programming recipe ($\boldsymbol{A} = \boldsymbol{U}\,\mathrm{diag}(\lambda_i)\,\boldsymbol{U}^T$,
squeezing, detectors, loss) and each application are laid out explicitly in the
main text. What pushes it above `low` is breadth and exoticism: it covers five
distinct applications, each with its own mode / photon-number profile, and the
primitives (hafnian, loop hafnian, Torontonian, GBS itself) are non-standard
photonic objects rather than gate-model circuits. Julien will need to
synthesise one fingerprint across a multi-application review rather than read a
single concrete experiment. No SI dependence and no exotic-fetch risk, hence
not `high`.