pith. sign in

arxiv: 2511.16945 · v1 · submitted 2025-11-21 · 🪐 quant-ph · cond-mat.dis-nn· physics.optics

Spectroscopy on a single nonlinear mode recognizes quantum states

Wouter Verstraelen , Stanis{\l}aw \'Swierczewski , Andrzej Opala , Andrew Haky , Matteo Gadani , Huawen Xu , Oleksandr Kyriienko , Micha{\l} Matuszewski
show 2 more authors
Alberto Bramati Timothy C.H. Liew
This is my paper

Pith reviewed 2026-05-17 21:17 UTC · model grok-4.3

classification 🪐 quant-ph cond-mat.dis-nnphysics.optics
keywords quantum reservoirsqueezed statesemission spectrumlinear regressionnonlinear modedriven-dissipativepolariton microcavityquantum state recognition
0
0 comments X

The pith

A single nonlinear driven-dissipative quantum mode can recognize parameters of incident squeezed states from its emission spectrum using linear regression.

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

The paper establishes that a quantum nonlinear driven-dissipative mode is sufficient to serve as a quantum reservoir for characterizing squeezed optical states. Instead of full quantum state tomography, which demands complex measurements, the approach extracts information from the occupations at different frequencies in the mode's emission spectrum and applies linear regression to recover the relevant state parameters in many cases. This holds under continuous driving and is shown explicitly when a degenerate optical parametric oscillator couples to a nonlinear polariton microcavity. A reader would care because the result points to a practical, hardware-light route to quantum state recognition that leverages existing nonlinear optical platforms.

Core claim

A quantum nonlinear driven-dissipative mode is sufficient to act as a quantum reservoir. By analyzing the occupations at different frequencies in the emission spectrum, a linear regression suffices in many cases to recognize the relevant parameters of incident squeezed states. The demonstration covers general continuous driving and a concrete example with a degenerate optical parametric oscillator coupled to a nonlinear polariton microcavity.

What carries the argument

Frequency-resolved occupations in the emission spectrum of the nonlinear mode, used as input features for linear regression to extract squeezed-state parameters.

If this is right

  • Linear regression on spectral occupations recovers squeezed-state parameters without requiring full tomography.
  • The method operates effectively under continuous driving of the nonlinear mode.
  • The approach succeeds for a source consisting of a degenerate optical parametric oscillator coupled to the nonlinear polariton microcavity.
  • This establishes a quantum reservoir based on a single driven-dissipative mode for state recognition tasks.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same spectral-occupation features might support recognition of other nonclassical states if the linear regressor is retrained accordingly.
  • Integration into compact photonic chips could lower the resource cost of routine quantum state checks compared with standard tomography setups.
  • Varying the nonlinearity strength or driving regime in follow-up experiments would map the boundary where linear regression ceases to suffice.
  • The reservoir idea connects naturally to tasks such as real-time feedback control of squeezed-light sources in quantum networks.

Load-bearing premise

The frequency-resolved occupations in the emission spectrum contain enough independent information about the incident squeezed state parameters for linear regression to recover them accurately.

What would settle it

Perform linear regression on measured frequency occupations from the emission spectrum of the driven nonlinear polariton microcavity with a degenerate OPO input; systematic failure to recover the squeezing parameters to within experimental error would falsify the claim.

Figures

Figures reproduced from arXiv: 2511.16945 by Alberto Bramati, Andrew Haky, Andrzej Opala, Huawen Xu, Matteo Gadani, Micha{\l} Matuszewski, Oleksandr Kyriienko, Stanis{\l}aw \'Swierczewski, Timothy C.H. Liew, Wouter Verstraelen.

Figure 1
Figure 1. Figure 1: Setup: Quantum-optical squeezed states are continuously injected in a single￾mode nonlinear cavity. The cavity’s output ra￾diation is then sent to a spectrometer, and the resulting parameters are then learned from the resulting emission spectrum by the Hamiltonian Hˆ = U 2 aˆ † aˆ † aˆa, ˆ (1) where aˆ is the reservoir node bosonic annihi￾lation operator1 . The evolution of a quantum system that is, under … view at source ↗
Figure 2
Figure 2. Figure 2: Learning squeezing strength (a): Spectra of a polariton microcavity coupled to a squeezed environment, for different values of squeezing strength r at θ = 0, αD = 5, n¯ = 0 (b): moments Mm as a function of r for the same training data as the left panel (circles) and 100 random testing data in the same range (stars). The moments are rescaled to their respective maximum values for display (c): individual com… view at source ↗
Figure 3
Figure 3. Figure 3: Dependence on reservoir non￾linearity (a): the prediction error on r (same task as in [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Generalisability of learning: rescaled moments Mm as function of r, when training is performed at θ = π/2 (circles) and testing at θ = 0 (stars). The prediction error on r is still relatively small at ε = 8.0%, highlight￾ing that learning in the network generalises. the digital computer used for post-processing). The number of neurons per layer was set as N1 = 200, N2 = 200, N3 = 100, and the out￾put layer… view at source ↗
Figure 5
Figure 5. Figure 5: Multivariate prediction with a hybrid quantum-classical model. In a dataset with variability both in r and θ, learning with moments M0 − M4 as before works well to predict r (inset of panel (a)), but less for θ (inset of (b)). Results improve significantly by training on the pre-processed (see panel c) full spectrum, as depicted on the main panels of (a,b). Further improvement is obtained using a feedforwa… view at source ↗
Figure 6
Figure 6. Figure 6: Spectroscopy reservoir ap￾proach, applied to a realistic OPO: a) Spectral differences of the polaritons driven by an OPO for different values of G with respect to the unsqueezed (G = 0) coher￾ently pumped spectrum S0, whole spectra in the inset. Spectra for test data G = 0 ms−1 , 0.5 ms−1 , . . . , 4.5 ms−1 are shown from purple to yellow. Moments are extracted from the spectral difference. Recognizing G o… view at source ↗
Figure 7
Figure 7. Figure 7: Error of predicting r (panel a) and θ (panel b) as a function of the highest spec￾tral moment (HSM) considered for regression. Within moments up to order m ≤ 4, we ob￾serve a reasonable convergence to the result of the full spectrum (gray, dashed). Ridge param￾eter 10−14 was used to avoid large deviations for the highest number of moments or the full spectrum. 4.3 OPO and polariton model We now move on to … view at source ↗
Figure 8
Figure 8. Figure 8: Dependence of the prediction error ε on the presence of external noise in the spec￾trum, during estimation of r. Whereas consid￾ering moments with mmax ≤ 2 always leads to a significant value of ε, moderate value of mmax give an optimal result when the noise becomes significant. been shown to be suitable for the description of polariton systems with many modes for the context of reservoir computing. 70 In … view at source ↗
read the original abstract

Characterising optical quantum states is essential for the development of quantum technologies. While traditional approaches to perform full quantum state tomography are often experimentally demanding, neuromorphic architectures may provide an effective alternative. In this work, we demonstrate how a quantum nonlinear driven-dissipative mode is sufficient to act as a quantum reservoir. By analyzing the occupations at different frequencies in the emission spectrum, a linear regression suffices in many cases to recognize the relevant parameters of incident squeezed states. Beyond highlighting the general potential of this approach under continuous driving, we illustrate its effectiveness in an explicit nontrivial example where the source is a degenerate optical parametric oscillator (OPO), coupled to a nonlinear polariton microcavity.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript claims that a single quantum nonlinear driven-dissipative mode is sufficient to function as a quantum reservoir for characterizing incident squeezed states. Under continuous driving, the frequency-resolved occupations in the steady-state emission spectrum provide a feature vector from which a linear regression can recover the amplitude and phase parameters of the squeezed state in many cases. The approach is illustrated with a concrete example of a degenerate optical parametric oscillator (OPO) coupled to a nonlinear polariton microcavity.

Significance. If the central claim is substantiated, the work offers a hardware-efficient alternative to full quantum state tomography that relies on a minimal physical system (one nonlinear mode) and standard spectral measurements rather than complex interferometric setups. It demonstrates the potential of reservoir-computing ideas in driven-dissipative quantum optics and provides an explicit, nontrivial example that could guide experiments with polariton or circuit-QED platforms. The continuous-driving regime is a practical strength.

major comments (2)
  1. [OPO-polariton example and linear-regression analysis] The central claim that linear regression on frequency occupations suffices rests on the assumption that these occupations furnish a feature vector whose linear span is rich enough to invert for the squeezed-state parameters. The skeptic note correctly identifies that, for perturbative nonlinearity or narrow-band drive, many frequency bins can become linearly dependent or insensitive to higher-order input correlations, rendering the design matrix ill-conditioned. The OPO-polariton example should therefore report the condition number (or singular-value spectrum) of the regression matrix for the chosen detuning and Kerr strength to demonstrate that the sidebands remain distinguishable; without this, it is unclear whether success is generic or an artifact of the specific parameters.
  2. [Results and discussion of regression performance] The abstract states that linear regression 'suffices in many cases' but provides no quantitative error analysis, success-rate statistics, or robustness checks against experimental noise or parameter variation. A load-bearing claim of this type requires explicit metrics (e.g., mean-squared error on recovered amplitude/phase or failure rate over a parameter sweep) in the results section to allow the reader to judge the regime of validity.
minor comments (2)
  1. Clarify the precise definition of 'occupations at different frequencies' (e.g., whether these are steady-state photon numbers in frequency bins of the output spectrum or integrated intensities) and how they are extracted from the emission spectrum.
  2. Add a brief comparison, even qualitative, to existing reservoir-computing or machine-learning approaches for quantum-state discrimination to better situate the novelty of using a single nonlinear mode.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments and positive overall assessment of our work. We address each major comment below and have revised the manuscript to strengthen the presentation of the linear-regression analysis.

read point-by-point responses

  1. Referee: [OPO-polariton example and linear-regression analysis] The central claim that linear regression on frequency occupations suffices rests on the assumption that these occupations furnish a feature vector whose linear span is rich enough to invert for the squeezed-state parameters. The skeptic note correctly identifies that, for perturbative nonlinearity or narrow-band drive, many frequency bins can become linearly dependent or insensitive to higher-order input correlations, rendering the design matrix ill-conditioned. The OPO-polariton example should therefore report the condition number (or singular-value spectrum) of the regression matrix for the chosen detuning and Kerr strength to demonstrate that the sidebands remain distinguishable; without this, it is unclear whether success is generic or an artifact of the specific parameters.

    Authors: We agree that explicitly reporting the conditioning of the regression matrix strengthens the claim. In the revised manuscript we have added the singular-value spectrum (and the resulting condition number) of the design matrix constructed from the frequency-resolved occupations for the OPO-polariton parameters used in the example. The spectrum confirms that the relevant sidebands remain linearly independent and that the matrix is sufficiently well-conditioned for stable inversion, indicating that the reported success is not an artifact of the chosen detuning and Kerr strength. revision: yes

  2. Referee: [Results and discussion of regression performance] The abstract states that linear regression 'suffices in many cases' but provides no quantitative error analysis, success-rate statistics, or robustness checks against experimental noise or parameter variation. A load-bearing claim of this type requires explicit metrics (e.g., mean-squared error on recovered amplitude/phase or failure rate over a parameter sweep) in the results section to allow the reader to judge the regime of validity.

    Authors: We acknowledge that the original manuscript would benefit from quantitative performance metrics. The revised results section now includes mean-squared errors on the recovered amplitude and phase, success rates (fraction of trials with relative error below a stated threshold) over a parameter sweep of squeezed-state amplitudes and phases, and robustness tests in which controlled Gaussian noise is added to the occupation features. These additions make the regime of validity explicit while preserving the central observation that linear regression works in many cases. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained via physical model and numerical demonstration

full rationale

The paper's central claim rests on modeling a driven-dissipative nonlinear mode, computing its steady-state emission spectrum for varied input squeezed-state parameters, and then applying linear regression to recover those parameters from frequency-resolved occupations. This workflow is a standard forward simulation followed by supervised regression on generated data; the regression coefficients are fitted to the model's outputs rather than presupposing the target recognition result. No self-definitional loops, fitted inputs relabeled as predictions, or load-bearing self-citations appear in the provided abstract or description. The approach remains falsifiable against external benchmarks or different nonlinearities, satisfying the criteria for an independent derivation chain.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard assumptions of driven-dissipative quantum optics (Markovian baths, coherent driving) and the sufficiency of spectral occupations for state discrimination. No explicit free parameters or invented entities are mentioned in the abstract.

axioms (1)
  • domain assumption The system is described by standard quantum optical master equations for a driven nonlinear mode coupled to a reservoir.
    Invoked implicitly to justify the emission spectrum analysis.

pith-pipeline@v0.9.0 · 5459 in / 1252 out tokens · 25463 ms · 2026-05-17T21:17:37.881670+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

  1. [1]

    D.; Kulkarni, S

    (2) Schuman, C. D.; Kulkarni, S. R.; Parsa, M.; Mitchell, J. P.; Date, P.; Kay, B. Opportunities for neuromorphic computing algorithms and applications. Nature Computational Science2022,2, 10–19. (3) Momeni,A.etal.Trainingofphysicalneu- ral networks.Nature2025,645, 53–61. (4) Shastri, B. J.; Tait, A. N.; Fer- reira de Lima, T.; Pernice, W. H. P.; Bhaskara...

    work page 2025

  2. [2]

    Role of all-optical neural networks.Phys

    (10) Matuszewski, M.; Prystupiuk, A.; Opala, A. Role of all-optical neural networks.Phys. Rev. Appl.2024,21, 014028. (11) McMahon, P. L. The physics of optical computing.Nature Reviews Physics2023, 5, 717–734. (12) Stroev, N.; Berloff, N. G. Analog Pho- tonics Computing for Information Pro- cessing, Inference, and Optimization.Ad- vanced Quantum Technolog...

    work page 2024

  3. [3]

    J.; Eisert, J.; Meyer, J

    (16) Schreiber, F. J.; Eisert, J.; Meyer, J. J. Tomography of Parametrized Quantum States.PRX Quantum2025,6, 020346. (17) Lvovsky, A. I.; Raymer, M. G. Continuous-variable optical quantum- state tomography.Rev. Mod. Phys.2009, 81, 299–332. (18) Ghosh, S.; Opala, A.; Matuszewski, M.; Paterek, T.; Liew, T. C. Quantum reser- voir processing.npj Quantum Infor...

    work page 2009

  4. [4]

    (19) Liu, Z.-H.; Brunel, R.; Østergaard, E. E. B.; Cordero, O.; Chen, S.; Wong, Y.; Nielsen, J. A. H.; Bregnsbo, A. B.; 15 Zhou, S.; Huang, H.-Y.; Oh, C.; Jiang, L.; Preskill, J.; Neergaard-Nielsen, J. S.; An- dersen, U. L. Quantum learning advantage on a scalable photonic platform.Science 2025,389, 1332–1335. (20) Pehle, C.; Wetterich, C. Neuromorphic qu...

    work page 2025

  5. [5]

    A.; Türeci, H

    (32) Angelatos, G.; Khan, S. A.; Türeci, H. E. Reservoir Computing Approach to Quan- tum State Measurement.Phys. Rev. X 2021,11, 041062. (33) der Sande, G. V.; Brunner, D.; Sori- ano, M. C. Advances in photonic reservoir computing.Nanophotonics2017,6, 561–

    work page 2021

  6. [6]

    (34) Xu, H.; Ghosh, S.; Matuszewski, M.; Liew, T. C. Universal Self-Correcting Computing with Disordered Exciton- Polariton Neural Networks.Phys. Rev. Appl.2020,13, 064074. (35) Xu, H.; Krisnanda, T.; Verstraelen, W.; Liew, T. C. H.; Ghosh, S. Superpolyno- mial quantum enhancement in polaritonic neuromorphic computing.Phys. Rev. B 2021,103, 195302. (36) D...

    work page 2020

  7. [7]

    L.; Girvin, S

    (38) Blais, A.; Grimsmo, A. L.; Girvin, S. M.; Wallraff, A. Circuit quantum electro- dynamics.Rev. Mod. Phys.2021,93, 025005. (39) Sheremet, A. S.; Petrov, M. I.; Iorsh, I. V.; Poshakinskiy, A. V.; Poddubny, A. N. 16 Waveguide quantum electrodynamics: Collective radiance and photon-photon correlations.Rev. Mod. Phys.2023,95, 015002. (40) Kavokin, A.; Liew...

    work page 2021

  8. [8]

    (43) Liew, T. C. H. The future of quantum in polariton systems: opinion.Opt. Mater. Express2023,13, 1938–1946. (44) Kalt, H.; Klingshirn, C. F.Semiconduc- tor optics 1: Linear Optical Properties of semiconductors; Springer International Publishing,

    work page 1938

  9. [9]

    Quantum fluids of light.Rev

    (55) Carusotto, I.; Ciuti, C. Quantum fluids of light.Rev. Mod. Phys.2013,85, 299–366. (56) Kasprzak, J.; André, R.; Dang, L. S.; She- lykh, I. A.; Kavokin, A. V.; Rubo, Y. G.; Kavokin, K. V.; Malpuech, G. Build up and pinning of linear polarization in the Bose condensates of exciton polaritons. Phys. Rev. B2007,75, 045326. (57) Liew, T. C. H.; Savona, V....

    work page 2013

  10. [10]

    Optical neural networks: progress and challenges.Light: Science & Applications 2024,13,

    (59) Fu, T.; Zhang, J.; Sun, R.; Huang, Y.; Xu, W.; Yang, S.; Zhu, Z.; Chen, H. Optical neural networks: progress and challenges.Light: Science & Applications 2024,13,

    work page 2024

  11. [11]

    (60) Govia, L. C. G.; Ribeill, G. J.; Row- lands, G. E.; Krovi, H. K.; Ohki, T. A. Quantum reservoir computing with a sin- gle nonlinear oscillator.Phys. Rev. Res. 2021,3, 013077. (61) Neill, O. D.; Nerenberg, S.; Marcucci, G.; Faccio, D. Quantum reservoir computing with linear photonic networks. Quantum Computing, Communication, and Simula- tion V. 2025;...

    work page 2021

  12. [12]

    (64) Carreño, J. C. L.; Laussy, F. P. Excita- tion with quantum light. I. Exciting a har- monic oscillator.Phys. Rev. A2016,94, 063825. (65) Gardiner, C. W. Driving a quantum sys- tem with the output field from another driven quantum system.Phys. Rev. Lett. 1993,70, 2269–2272. (66) Carmichael, H. J. Quantum trajectory theory for cascaded open systems.Phys...

    work page 1993

  13. [13]

    J.; Lewenstein, M

    (69) Glauber, R. J.; Lewenstein, M. Quantum optics of dielectric media.Phys. Rev. A 1991,43, 467–491. (70) Świerczewski, S.; Verstraelen, W.; Deuar, P.; Piętka, B.; Liew, T. C. H.; Matuszewski, M.; Opala, A. Phase-Space Framework for Noisy Intermediate-Scale Quantum Optical Neural Networks

    work page 1991

  14. [14]

    A.; Salem, R.; Okawachi, Y.; Turner-Foster, A

    (74) Foster, M. A.; Salem, R.; Okawachi, Y.; Turner-Foster, A. C.; Lipson, M.; Gaeta, A. L. Ultrafast waveform com- pression using a time-domain telescope. Nature Photonics2009,3, 581–585. (75) Joshi, C.; Sparkes, B. M.; Farsi, A.; Gerrits, T.; Verma, V.; Ramelow, S.; Nam, S. W.; Gaeta, A. L. Picosecond- resolution single-photon time lens for tem- poral m...

    work page 2022