Training the parametric interactions in an analog bosonic quantum neural network with Fock basis measurement

Baptiste Carles; Danijela Markovi\'c; Elie Gouzien; Julie Grollier; Julien Dudas

REVIEW 3 major objections 5 minor 28 references

Training the parametric interactions in an analog bosonic quantum neural network with Fock basis measurement

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Physical parameters of a linear bosonic network can be trained by backpropagating through a classical Gaussian simulation, letting a small network classify with far fewer measurements than an untrained reservoir.

desk verdict A genuinely useful and honest numerical study of trainable linear bosonic QNNs, whose central claim holds for the tested small-mode regimes but whose broader scalability case is weakened by a scaling contradiction and stability heuristics the authors admit are unproven. read the letter →

arxiv 2411.19112 v3 submitted 2024-11-28 quant-ph cond-mat.dis-nn

Julien Dudas , Baptiste Carles , Elie Gouzien , Julie Grollier , Danijela Marković This is my paper · ORCID

classification quant-phcond-mat.dis-nn

keywords bosonicquantumneuralnetworkGaussianstatesFockbasismeasurementbackpropagationbosonsamplingreservoircomputingparametriccouplingcircuitQED

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that the physical parameters of a linear network of coupled bosonic modes - the drive amplitudes, phases, detunings, photon-exchange rates, and two-mode squeezing rates - can be trained cohesively by backpropagating through a classical simulation of Gaussian dynamics, with nonlinearity supplied by Fock-basis measurements. The authors demonstrate on three benchmark tasks that such end-to-end training lets a compact network of two to six modes reach or exceed the accuracy of untrained quantum reservoir computing while measuring far fewer outputs: a single measured probability suffices for the sine/square and spirals tasks, versus nine and thirty-six observables for the reservoir. They argue this hybrid scheme is practical because gradients come from the simulator while the forward evolution could run on quantum hardware, avoiding parameter-shift rules or device-level gradient extraction. If correct, trainable analog quantum neural networks can be designed and optimized before hardware experiments.

What carries the argument

The central object is the quadratic Hamiltonian $H_0 = -\sum_k \delta_k a_k^\dagger a_k + \sum_{k<l}(g_{kl} a_k^\dagger a_l + g^s_{kl} a_k^\dagger a_l^\dagger + h.c.)$, whose coefficients are the trainable physical parameters. Its Gaussian dynamics give closed-form displacement and covariance matrices under dissipation, and Fock-basis probabilities $P_k(n)$ are computed by the Gaussian boson sampling formula involving the loop hafnian of a matrix built from the covariance matrix. This makes every output probability a differentiable function of the physical parameters, allowing automatic differentiation of the classical simulation to yield exact gradients. The choice of encoding parameter also matters: encoding in the two-mode squeezing rates outperforms other encodings because squeezing directly changes the covariance matrix, whereas photon exchange alone leaves it constant.

What would settle it

Train the same bosonic QNN on a larger benchmark (ten or more modes) using the published clamping rules and monitor average photon number during Adam iterations; a run in which photon numbers diverge before accuracy saturates, or in which clamping pins the encoder so the input no longer changes the dynamics, would show the training pipeline is not as scalable as claimed.

Watch

Extended reading notes

Core claim

The central discovery is that a quadratic Hamiltonian with parametric couplings, whose evolution generates Gaussian states, is not a limitation for quantum machine learning if the nonlinearity is placed in the measurement. Because Gaussian evolution is classically simulable in the Heisenberg picture, every Fock-basis output probability and its gradient with respect to the physical parameters can be computed analytically via Gaussian boson sampling and loop hafnians, so the whole forward map is differentiable end-to-end. Training is then ordinary Adam descent on a classical surrogate, while the same unitary evolution could be executed on a circuit-QED or integrated-photonic device. The benchmarks show this works for sine/square classification with two modes, spirals classification with four modes, and handwritten-digit recognition with six modes and data re-uploading, in each case using fewer measured observables than the quantum reservoir baseline and, for digits, achieving accuracy the reservoir cannot approach.

Load-bearing premise

The training procedure presumes that heuristic clamping rules (for example, two-mode squeezing amplitudes never exceeding the smallest photon-conversion rate) keep the dynamics inside a parameter region where photon numbers stay finite; the paper itself states these are practical intuitions, not proven guarantees, so if a gradient step escapes that region the training run diverges and the approach fails.

Editorial extensions

If this is right

Trained physical parameters can be optimized entirely in simulation and then transferred to hardware, since the forward evolution is the same quadratic dynamics the simulation models.
A single Fock-basis probability can serve as the only output feature for some classification tasks, cutting the number of observables from nine or thirty-six to one and the shot count from 10^8 to 10^3.
Training the drive and coupling parameters unlocks tasks that the untrained reservoir cannot solve: on the DIGITS dataset the trained network exceeds 97% accuracy while the reservoir tops out near 12%.
The choice of encoding parameter is itself a hyperparameter: encoding in two-mode squeezing amplitudes or phases reaches 100% spirals accuracy with one measured probability, while drive-phase or drive-amplitude encodings need more outputs and photon-exchange encoding plateaus at 85%.
Gradient-based training is feasible without parameter-shift rules or device-level gradient extraction, because the Gaussian simulation provides analytic gradients for the entire physical parameter set.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The same end-to-end training should transfer to any platform realizing the same quadratic Hamiltonian, including photonic squeezing lattices where trainable parameters are phase-matching conditions; this could be tested by porting trained parameters to a photonic chip and checking accuracy.
The strong performance of squeezing-based encoding suggests that to make a bosonic network trainable for classification, the encoding should modulate the covariance matrix rather than only displacements.
The authors conjecture but do not prove that bosonic networks avoid barren plateaus; if true, a testable corollary is that gradient variances stay bounded as the number of modes grows, which could be checked for ten-to-fifty-mode networks.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

This paper is worth a serious look. It proposes a trainable analog bosonic QNN where physical parameters—drive amplitudes, phases, detunings, photon-exchange and two-mode squeezing rates—are optimized by backpropagating through a classical Gaussian simulation, with Fock-basis measurement supplying the nonlinearity. The specific demonstrations are new: single-measurement sine/square and spirals classification, over 97% accuracy on DIGITS, and a large reduction in measured observables compared to an untrained quantum reservoir. The supplement includes a careful analysis of photon-number divergence regimes and a derivation of loop-hafnian gradients. The authors are also transparent: they explicitly state that the clamping rules used to keep training stable are practical intuitions, not proven methods. That honesty is a strength, not a weakness.

The main soft spot is the one the stress-test flags, and I think it is the real issue. The training pipeline relies on heuristic clamping after every gradient step to keep the system in a finite-photon region. No bound is given on the probability that an unclamped update crosses into divergence, and the supplement shows that divergence regions are generic for certain phase configurations. This makes the scalability claim conditional: the method is demonstrated for the small M and manually selected ranges tested, not for the general parameter family. That is a genuine limitation, but it does not undermine the central demonstration in the small-mode regime. The second concrete problem is internal: the abstract says trainable parameters scale only linearly with the number of modes, while the main text correctly gives a quadratic expression, 3/2 M(M−1)+3M. That mismatch needs fixing. Third, there is no public code, and apart from one encoding scheme averaged over initializations, no error bars are given for the QNN results, so independent reproduction is harder than it should be.

The comparison against reservoir computing is fair but narrow: it shows the advantage over an untrained network, but not against other trained continuous-variable QNNs. The self-citation is minor and the cited prior reservoir work is directly relevant.

Who benefits: researchers in quantum reservoir computing, continuous-variable QML, and experimentalists in cQED or integrated photonics who want a trainable architecture compatible with parametric couplers. I would send this to referees; the scaling contradiction, the stability heuristics, and the missing code are addresses in revision, and the core trainability result is solid enough to warrant that effort. I would cite it if I worked in this area.

Referee Report

3 major / 5 minor

Summary. The paper proposes an analog bosonic quantum neural network (QNN) built from parametrically coupled Gaussian modes, with nonlinearity supplied by Fock-basis measurements. The authors derive the Gaussian dynamics in the Heisenberg picture, compute Fock-state probabilities via the Gaussian boson sampling formula, and use automatic differentiation to train the physical parameters (drive amplitudes, detunings, photon-exchange and two-mode squeezing rates) together with the output weights. They benchmark the architecture on three classification tasks (sine/square, spirals, DIGITS) and report that the trained network reaches high accuracy with dramatically fewer measured observables and measurement shots than an untrained quantum reservoir. The central claim is that these physically meaningful parameters can be trained cohesively and that the architecture is a scalable, experimentally realizable route to quantum machine learning.

Significance. If the central trainability claim holds, the work offers a practical hybrid classical-quantum training scheme for analog bosonic networks that is compatible with cQED parametric couplers and g(2)/g(3) photonic platforms, with the notable advantage of not requiring gradient readout from the hardware. The analytical gradient derivation, the explicit treatment of the Gaussian Langevin dynamics, and the reproducible small-scale benchmarks are strengths. However, the generality of the claim is weakened by two load-bearing issues: an internal contradiction between the abstract and the main text on parameter scaling, and the explicit caveat in the supplement that the divergence-prevention clamping rules are unproven heuristics. These issues make the current demonstration conditional on hand-constrained parameter regimes and small mode numbers rather than a general scalability statement.

major comments (3)

[Abstract vs. Section 2 (around Eq. (2))] The abstract states that 'the number of trainable parameters scales only linearly with the number of modes', but the main text gives the number of trainable physical parameters as 3M(M−1)/2 + 3M, which is quadratic in M. This is not a typographical subtlety: it changes the advertised scalability of the architecture. The authors must either correct the abstract, or clarify what quantity is meant to scale linearly (e.g., number of measured outputs or number of squeezing tones per mode) and reconcile the text.
[Supplementary Section VII, Algorithm 2] The stability of training rests on a set of clamping rules applied after every gradient update. The supplement explicitly says these 'are not rigorously proven methods for preventing such divergences, but rather practical intuitions that have been effective in training the coupling parameters' (Supp. Sec. VII). Since Supplementary Section VI demonstrates that generic phase configurations of the couplings lead to exponential photon-number divergence, the clamping rules are not a peripheral detail but the mechanism that keeps the simulation finite. The central clause 'we demonstrate that these parameters ... can be trained cohesively' is therefore demonstrated only for the hand-picked parameter ranges and small mode counts tested, not for the general parametric family. A rigorous stability bound, or at least a quantitative characterization of the region in which the clamps are active, is needed to support the claimed scalability.
[Supplementary Section V, 'Derivatives with respect to eigenvectors'] The gradient computation through eigendecomposition is only well-defined when F' has non-degenerate eigenvalues, and the authors state that initial parameters are chosen accordingly. However, training may drive the parameters toward degeneracy — the supplement itself identifies the degeneracy condition |g| = |g_s| in the two-mode case — and no mechanism or check is described to ensure that the non-degeneracy condition persists during optimization. If near-degeneracy is encountered, the training gradients themselves become unstable, which is a second, independent failure mode distinct from photon-number divergence. The authors should either add a safeguard in the optimization loop or provide evidence that the training trajectories used in the benchmarks remain safely away from degeneracies.

minor comments (5)

[Abstract and Section 1] The phrase 'extraction of an exponential number of features relative to the number of modes' is imprecise: the Fock basis for each mode is infinite, and in practice the features are truncated to a finite set of photon numbers. Clarify what exponential refers to (e.g., number of Fock basis states up to a fixed cutoff in a single mode).
[Eq. (4) and surrounding notation] The identity matrix is denoted '12' in Eq. (4), which is easily mistaken for a twelve-dimensional identity. Use a standard notation such as I_2 or 1_2, and likewise for the zero matrix.
[Table I and Section 2 (sine/square task)] The comparison with quantum reservoir computing reports 1000 shots for the bosonic QNN versus 10^8 for the reservoir, but it would be helpful to state explicitly that the 1000 shots refer only to the single measured probability P1(0), and whether the same accuracy is achieved with that shot count in a single run or after averaging over runs.
[Supplementary Section I, Table S1] The learning rates for δ, ϵ, g, and g_s are listed as 0.1 for some tasks and 0.01 for others, but the text does not explain whether these values were tuned per task or chosen by a common heuristic. A sentence on hyperparameter selection would aid reproducibility.
[Main text, Conclusion] The claim that bosonic neural networks 'may be inherently less susceptible to barren plateaus' is speculative and not supported by the benchmarks in the paper. This is fine as a future outlook, but it should be explicitly flagged as a conjecture rather than presented as a consequence of the results.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the training pipeline is a numerical experiment built on the external GBS formalism, and the self-cited reservoir baseline is not load-bearing.

full rationale

The paper's forward model uses the standard Gaussian Langevin equation and Gaussian boson sampling formula (Supp. Eqs. S5-S16), attributed to external references [15,16]; gradients are obtained by automatic differentiation of the same GBS expression, which is a consistency requirement rather than a circular reduction. The central claim that parameters with different physical dimensions can be trained cohesively is an empirical demonstration on benchmark tasks, not a derivation whose conclusion is assumed in its inputs. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work, and no ansatz is smuggled in via self-citation. The comparison to quantum reservoir computing uses the authors' own earlier work [13] as a baseline; this is a self-citation, but it is a separate published benchmark with its own fitted values and is external to the present training runs, so per the rubric it does not raise the circularity score. The clamping heuristics in Supplementary Section VII are explicitly acknowledged as not rigorously proven; this is a limitation on the generality of the trainability claim, but it is not a case of a result being equivalent to its inputs. No circular step meeting the quoted-evidence standard is present.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new physical entities. Its central claim rests on standard Gaussian quantum optics formulas, a linear-loss Langevin model, and two ad hoc stability assumptions: the heuristic clamping rules and the non-degeneracy of the eigenvector basis used for autodiff. The listed free parameters are training hyperparameters and clamping bounds that are chosen by hand and affect whether the demonstrated training succeeds.

free parameters (3)

regularization prefactor beta_<N> = 0.02 (sine/square, spirals), 0.12 (DIGITS)
Introduced in the loss to penalize average photon number and prevent divergence; values are chosen per task in Table S1, not derived.
target average photon number <N>_tg = 2 photons per mode
Loss target for the photon-number regularization; set by hand in the main text.
clamping bound l_g_max = 500 MHz, with related l_gs bounds and min(g)/(M-1) limits
Heuristic upper and lower bounds in Algorithm 2 to prevent destructive interference and photon divergence; the authors describe them as practical intuitions, not proven rules.

assumptions (6)

standard math Gaussian boson sampling formula for Fock probabilities (Eq. 3 and Supp. II C)
Used to compute output features and their parameter gradients; cited to Hamilton et al. [15].
standard math Heisenberg-Langevin solution for Gaussian states, including propagator exponentials and covariance integrals (Supp. II A, II B)
Provides the classical model of the dynamics through which gradients are backpropagated.
domain assumption Rotating-wave approximation with coupling-tone detunings fixed to (delta_k +/- delta_l)/2, and input modes in coherent states with constant amplitudes
Idealizes the hardware; real parametric couplers have residual and higher-order terms not modeled here.
ad hoc to paper Clamping rules in Supp. VII prevent photon-number divergence
The authors explicitly state these rules are not rigorously proven; if they fail, training becomes unstable and the central claim is unsupported.
ad hoc to paper Initial physical parameters are chosen so the matrix F' has non-degenerate eigenvalues
PyTorch's eigenvector gradients are only well defined for distinct eigenvalues (Supp. V); the training procedure depends on this condition.
domain assumption Single-mode marginal Fock probabilities P_k(n) are sufficient and the traced-out state remains Gaussian
Used to reduce the readout to a small number of measured observables; supports the measurement reduction claims.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Training the parametric interactions in an analog bosonic quantum neural network with Fock basis measurement." pith.science (2026). https://pith.science/paper/2411.19112

@misc{pith2026241119112,
  author       = {Pith},
  title        = {Pith review of: Training the parametric interactions in an analog bosonic quantum neural network with Fock basis measurement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2411.19112}},
  note         = {Machine review of arXiv:2411.19112}
}

abstract

Quantum neural networks promise to extend the power of machine learning into the quantum domain, with potential applications ranging from automatic recognition of quantum states to the control of quantum devices. However, their physical implementation and training remain challenging. In particular, the backpropagation algorithm that underpins the efficiency of classical neural networks cannot generally be applied to large quantum systems, as nonlinear quantum dynamics are not efficiently simulable. Instead, variational quantum circuits typically rely on parameter-shift rules or sampling-based gradient estimation. Here we propose a bosonic quantum neural network based on parametrically coupled Gaussian modes. Although the underlying quantum dynamics are linear, nonlinear output features are generated through Fock-basis measurements. Because Gaussian evolution can be efficiently simulated in the Heisenberg representation, the system admits gradient-based optimization by differentiating a classical model of the dynamics, while the forward evolution itself could be implemented on quantum hardware. This hybrid approach enables end-to-end training of physically meaningful parameters without requiring gradient extraction from the experimental device. Such architectures are naturally compatible with circuit quantum electrodynamics platforms featuring tunable parametric couplers, as well as integrated photonic systems with engineered $\chi$(2) or $\chi$(3) nonlinearities. Our results demonstrate that linear bosonic networks combined with nonlinear measurement provide a scalable and trainable route toward experimentally realizable quantum neural networks.

Figures

Figures reproduced from arXiv: 2411.19112 by the authors.

**Figure 2.** FIG. 2: Sine/square classification task. (a) The input [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** The input data for this task is two-dimensional, [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 3.** Figure 3: FIG. 3: The spirals classification task consists in as [PITH_FULL_IMAGE:figures/full_fig_p004_3.png]

**Figure 4.** Figure 4: FIG. 4: (a) A sample from the DIGITS dataset, con [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 21 canonical work pages

[1]

L. K. Grover, in Proceedings of the Twenty-Eighth Annual ACM Symposium on Theory of Computing , STOC ’96 (Association for Computing Machinery, New York, NY, USA, 1996) pp. 212–219

work page 1996
[2]

P. W. Shor, SIAM J. Comput. 26, 1484 (1997)

1997
[3]

A. W. Harrow, A. Hassidim, and S. Lloyd, Phys. Rev. Lett. 103, 150502 (2009)

2009
[4]

Benedetti, E

M. Benedetti, E. Lloyd, S. Sack, and M. Fiorentini, Quantum Sci. Technol. 4, 043001 (2019)

2019
[5]

X. Ding, Z. Song, J. Xu, Y. Hou, T. Yang, and Z. Shan, Sci Rep 14, 15886 (2024)

work page 2024
[6]

Herrmann, S

J. Herrmann, S. M. Llima, A. Remm, P. Zapletal, N. A. McMahon, C. Scarato, F. Swiadek, C. K. Andersen, C. Hellings, S. Krinner, N. Lacroix, S. Lazar, M. Ker- schbaum, D. C. Zanuz, G. J. Norris, M. J. Hartmann, A. Wallraff, and C. Eichler, Nat Commun 13, 4144 (2022)

work page 2022
[7]

Huang, M

H.-Y. Huang, M. Broughton, J. Cotler, S. Chen, J. Li, M. Mohseni, H. Neven, R. Babbush, R. Kueng, J. Preskill, and J. R. McClean, Science 376, 1182 (2022)

work page 2022
[8]

Benedetti, D

M. Benedetti, D. Garcia-Pintos, O. Perdomo, V. Leyton- Ortega, Y. Nam, and A. Perdomo-Ortiz, npj Quantum Inf 5, 1 (2019)

work page 2019

Show all 28 references

[9]

Hu, S.-H

L. Hu, S.-H. Wu, W. Cai, Y. Ma, X. Mu, Y. Xu, H. Wang, Y. Song, D.-L. Deng, C.-L. Zou, and L. Sun, Science Advances 5, eaav2761 (2019)

2019
[10]

A. V. Uvarov and J. D. Biamonte, J. Phys. A: Math. Theor. 54, 245301 (2021)

2021
[11]

Angelatos, S

G. Angelatos, S. A. Khan, and H. E. T¨ ureci, Phys. Rev. X 11, 041062 (2021)

2021
[12]

Senanian, S

A. Senanian, S. Prabhu, V. Kremenetski, S. Roy, Y. Cao, J. Kline, T. Onodera, L. G. Wright, X. Wu, V. Fatemi, and P. L. McMahon, Nat Commun 15, 7490 (2024)

2024
[13]

Dudas, B

J. Dudas, B. Carles, E. Plouet, F. A. Mizrahi, J. Grollier, and D. Markovi´ c, npj Quantum Inf9, 1 (2023)

2023
[14]

Metelmann, SciPost Physics Lecture Notes , 066 (2023)

A. Metelmann, SciPost Physics Lecture Notes , 066 (2023)

2023
[16]

Bj¨ orklund, B

A. Bj¨ orklund, B. Gupt, and N. Quesada, A faster hafnian formula for complex matrices and its benchmarking on a supercomputer (2019), arXiv:1805.12498 [quant-ph]

2019 arXiv
[17]

D. P. Kingma and J. Ba, Adam: A Method for Stochastic Optimization (2017), arXiv:1412.6980 [cs]

2017 arXiv
[18]

J. F. F. Bulmer, B. A. Bell, R. S. Chadwick, A. E. Jones, D. Moise, A. Rigazzi, J. Thorbecke, U.-U. Haus, 6 T. Van Vaerenbergh, R. B. Patel, I. A. Walmsley, and A. Laing, Science Advances 8, eabl9236 (2022)

2022
[19]

S. A. Khan, F. Hu, G. Angelatos, and H. E. T¨ ureci, Phys- ical reservoir computing using finitely-sampled quantum systems (2021), arXiv:2110.13849 [quant-ph]

2021 arXiv
[20]

P´ erez-Salinas, A

A. P´ erez-Salinas, A. Cervera-Lierta, E. Gil-Fuster, and J. I. Latorre, Quantum 4, 226 (2020). Supplementary Material for ”Training the parametric interactions in an analog bosonic quantum neural network with Fock basis measurement” J. Dudas, 1 B. Carles, 1 J. Grollier, 1 E. ...

2020
[21]

DIGITS classification task 4 II

Multi-Layer Perceptron 3 C. DIGITS classification task 4 II. Quantum Langevin equation and Gaussian Boson Sampling 4 A. Solution of the quantum Langevin equation 4 B. Computation of the displacement and covariance matrix of ladder operators α(t),σ(t) via diagonalization 5 C. G...

2025 arXiv
[22]

Since a single-layer perceptron cannot solve this problem, we employ an MLP with two hidden layers

Multi-Layer Perceptron To provide a classical point of comparison for our bosonic QNN, we compare its performance on the spirals classifi- cation task to a Multi-Layer Perceptron (MLP). Since a single-layer perceptron cannot solve this problem, we employ an MLP with two hidden...
[23]

(S26) Its diagonalization F′ =UΛU−1 yields Λ = diag(λ+,λ +,λ−,λ−) (S27) U = 1 N   i(|gs|2−|g|2) −g i (|gs|2−|g|2) −g g∗λ+ −iλ+ g∗λ− −iλ− 0 gs∗ 0 gs∗ −gs∗λ+ 0 −gs∗λ− 0  , (S28) 9 withN a normalization factor and λg,gs = { −i √ |g|2−|gs|2 if|g|>|gs|√ |gs|2−|g|2 if|gs|>|g| ...
[24]

By symmetry, the mean photon numbers in mode 1 and 2 are equal: ⟨N1⟩ =⟨N2⟩ =⟨N⟩

(S30) If|g| =|gs| i.e the photon conversion rate and two-mode squeezing rate have equal absolute values, then F′ is not diagonalizable. By symmetry, the mean photon numbers in mode 1 and 2 are equal: ⟨N1⟩ =⟨N2⟩ =⟨N⟩. Then from Eq. (S22) we get ⟨N(t)⟩ =κ|ϵ|2 2M∑ k′,m′,k,m=1 U1+...
[25]

We observe that for certain phase values the number of photons diverges. We interpret this as a consequence of the destructive interference between multiple photon conversion processes, resulting in the creation of photons through two-mode squeezing processes, and the lack of ...
[26]

Gouzien, Optique quantique multimode pour le traitement de l’information quantique , Ph.D

´E. Gouzien, Optique quantique multimode pour le traitement de l’information quantique , Ph.D. thesis, COMUE Universit´ e Cˆ ote d’Azur (2015 - 2019) (2019)

2019
[27]

C. S. Hamilton, R. Kruse, L. Sansoni, S. Barkhofen, C. Silberhorn, and I. Jex, Phys. Rev. Lett. 119, 170501 (2017)

2017
[28]

Torch.linalg.eig — PyTorch 2.6 documentation (2024)

2024
[29]

Banchi, N

L. Banchi, N. Quesada, and J. M. Arrazola, Phys. Rev. A 102, 012417 (2020)

2020

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

L. K. Grover, in Proceedings of the Twenty-Eighth Annual ACM Symposium on Theory of Computing , STOC ’96 (Association for Computing Machinery, New York, NY, USA, 1996) pp. 212–219

work page 1996

[2] [2]

P. W. Shor, SIAM J. Comput. 26, 1484 (1997)

1997

[3] [3]

A. W. Harrow, A. Hassidim, and S. Lloyd, Phys. Rev. Lett. 103, 150502 (2009)

2009

[4] [4]

Benedetti, E

M. Benedetti, E. Lloyd, S. Sack, and M. Fiorentini, Quantum Sci. Technol. 4, 043001 (2019)

2019

[5] [5]

X. Ding, Z. Song, J. Xu, Y. Hou, T. Yang, and Z. Shan, Sci Rep 14, 15886 (2024)

work page 2024

[6] [6]

Herrmann, S

J. Herrmann, S. M. Llima, A. Remm, P. Zapletal, N. A. McMahon, C. Scarato, F. Swiadek, C. K. Andersen, C. Hellings, S. Krinner, N. Lacroix, S. Lazar, M. Ker- schbaum, D. C. Zanuz, G. J. Norris, M. J. Hartmann, A. Wallraff, and C. Eichler, Nat Commun 13, 4144 (2022)

work page 2022

[7] [7]

Huang, M

H.-Y. Huang, M. Broughton, J. Cotler, S. Chen, J. Li, M. Mohseni, H. Neven, R. Babbush, R. Kueng, J. Preskill, and J. R. McClean, Science 376, 1182 (2022)

work page 2022

[8] [8]

Benedetti, D

M. Benedetti, D. Garcia-Pintos, O. Perdomo, V. Leyton- Ortega, Y. Nam, and A. Perdomo-Ortiz, npj Quantum Inf 5, 1 (2019)

work page 2019