Pith. sign in

REVIEW 3 major objections 5 minor 67 references

Dissipation alters modes of information encoding in small quantum reservoirs near criticality

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper claims that near a critical coupling, a driven-dissipative pair of Kerr oscillators crosses from redundant to synergistic encoding, with photon loss tuning the balance and the associated memory profile.

desk verdict The empirical redundant-to-synergistic transition near J=|Δ| looks real across three levels of simulation, but the Keldysh-pole mechanism in Sec. 3.2 is computed at the undriven α=0 fixed point and does not describe the driven regime where the peak appears. read the letter →

arxiv 2412.18290 v2 pith:GRLT4NEH submitted 2024-12-24 quant-ph cond-mat.dis-nncond-mat.stat-mechcs.LG

Krai Cheamsawat , Thiparat Chotibut This is my paper · ORCID
classification quant-phcond-mat.dis-nncond-mat.stat-mechcs.LG MSC 81P4581V80
keywords quantumreservoircomputingcoupledKerroscillatorspartialinformationdecompositionsynergyredundancydissipationdynamicalbifurcationmemorycapacity
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

This paper tries to establish that the information-encoding style of a small quantum reservoir can be dialed by two knobs: coupling and photon loss. The system is a pair of coupled Kerr-nonlinear oscillators, two light modes that interact nonlinearly and leak photons at a tunable rate, driven by one common signal. Near the critical coupling $J=|\Delta|$, where a slow collective mode becomes overdamped, the reservoir switches from storing duplicate information about the input (redundancy) to storing information that appears only when both oscillators are read together (synergy). The paper further claims that synergy sharpens short-term memory while strong dissipation restores redundancy and supports long-term retention. If right, this gives a concrete design principle: choose the coupling-to-detuning ratio and photon-loss rate to select the reservoir's memory profile.

What carries the argument

The load-bearing object is the retarded Green's function $G^R(\omega)$ of the Keldysh mean-field fluctuation action. Its poles come in two branches, $$\omega_s = \pm\bigl|J-|\$\Delta$|\bigr| - i\gamma, \qquad \omega_f = \pm\bigl|J+|\$\Delta$|\bigr| - i\gamma,$$ so at $J=|\Delta|$ the slow branch loses its real part and becomes purely decaying, while the fast branch remains oscillatory. The paper interprets this as an overdamped soft mode: the would-be flat direction of the effective potential becomes a non-oscillatory relaxation channel, and the response is dominated by a single coherent fast oscillation. This single-timescale coherence is what the paper identifies as the mechanism that reduces overlap between modes and makes information jointly accessible only through both readouts, i.e., synergistic encoding.

What would settle it

Compute the retarded Green's function poles linearized around the actual time-averaged driven steady state for $F=0.2$ and $F=2.0$ by solving the mean-field equations and inserting the displaced amplitudes into the inverse Green's function of Eq. (C.4). If the slow-mode real part does not vanish near $J=|\Delta|$ under the driven steady state, the proposed soft-mode mechanism would not explain the synergy peak; alternatively, a sweep of the synergy peak versus $J$ at increasing drive strength should show whether the peak tracks $J=|\Delta|$ or moves with the driven bifurcation.

Watch

Extended reading notes

Core claim

The paper's central claim is that in a pair of coupled Kerr-nonlinear oscillators driven by a common time-dependent signal, the way the two measured outputs encode the signal is controlled by proximity to a dynamical bifurcation and by photon loss. Using partial information decomposition on the readouts $X_i = \mathrm{Re}\langle \hat a_i\rangle$, the authors find that at coupling $J$ equal to the absolute detuning $|\Delta|$, normalized synergy peaks while redundancy drops: the joint readout contains information about the drive that neither oscillator alone carries. They attribute the peak to a linear-response mechanism: at $J=|\Delta|$ the effective potential around the zero steady state develops flat directions, and with dissipation the corresponding slow modes become overdamped, leaving the response dominated by one coherent fast oscillation. This coherence favors collective, synergistic encoding. Increasing the photon-loss rate $\gamma$ overdamps both modes, makes the two oscillators nearly identical and nearly independent, and shifts encoding back to redundancy. The paper then connects these modes to memory: synergy sharpens short-term response, redundancy stabilizes long-term retention, and at high $\gamma$ the total memory capacity at criticality grows because rapid relaxation reduces output variance even though the correlation decay rate stays roughly constant.

Load-bearing premise

The load-bearing premise is that the linear-response pole structure computed around the undriven steady state $\alpha_{1,c}=\alpha_{2,c}=0$, as in Eqs. (12)-(13) and Eq. (C.7), describes the encoding dynamics when the drive $F(t)$ is present; if the driven steady state's slow and fast modes differ, the overdamped-soft-mode mechanism fails, even though the measured synergy peak could still exist.

Editorial extensions

If this is right

  • Near $J=|\Delta|$, normalized synergy peaks and normalized redundancy drops in mean-field, second-order cumulant, and fully quantum simulations, so the redundant-to-synergistic crossover is not an artifact of one approximation.
  • The same crossover appears for telegraph and uniform uncorrelated inputs, meaning the effect is a property of the reservoir's response dynamics rather than of the input statistics.
  • Stronger photon loss $\gamma$ lowers quantum mutual information and synergy, pushes the two oscillators toward a product state, and makes encoding predominantly redundant.
  • At critical coupling, low dissipation favors short-delay memory while long-delay capacity decays quickly; higher dissipation makes total memory capacity grow, with the correlation decay exponent roughly independent of $\gamma$.
  • The second-order cumulant expansion interpolates between mean-field and full quantum results, placing partial quantum correlations at an intermediate encoding bias.

Reading between the lines

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

  • Beyond the paper: the mechanism's predictions could be probed at larger drive amplitudes, where the steady state is displaced; if the synergy peak follows the driven steady-state bifurcation instead of the undriven $J=|\Delta|$ condition, the linear-response story would need revision.
  • Beyond the paper: similar soft-mode-overdamping arguments may predict synergy peaks at dynamical bifurcations in larger driven-dissipative lattices, making the crossover a generic feature of collective instabilities rather than a two-oscillator specialty.
  • Beyond the paper: since the PID analysis uses classical readouts, a quantum analogue of synergy defined on the joint density matrix could reveal whether the synergy near criticality is accompanied by genuine nonclassical resources or only by classically correlated oscillations.
  • Beyond the paper: the memory-capacity results suggest a practical tuning rule for reservoir design: operate near criticality with weak dissipation for tasks needing fast reaction to recent inputs, and away from criticality with stronger dissipation for tasks needing long stable memory.
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.

Referee Report

3 major / 5 minor

Summary. The paper studies two coupled, driven-dissipative Kerr-nonlinear oscillators as a minimal quantum reservoir. The input signal s(t) drives both oscillators, and the readouts are the mean-field quadratures X_i=Re<â_i>. Using Partial Information Decomposition (PID) with the BROJA prescription, the authors separate the mutual information between the input and the joint readout into redundant, synergistic, and unique components. They report a peak in normalized synergy near the coupling-detuning resonance J=|Δ|, observed consistently in mean-field, second-order cumulant, and full quantum master-equation simulations, and robust to both telegraph and uniform-random input signals. They attribute this peak to the overdamping of a soft collective mode and the resulting dominance of coherent fast oscillations, using a Keldysh linear-response calculation of the retarded Green's function. They then examine how increasing photon loss γ reduces quantum correlations, biases encoding toward redundancy, and affects short-term memory capacity in a linear-readout reservoir benchmark.

Significance. If the central mechanism were established, the paper would provide a concrete design principle for small quantum reservoirs: operating near a dynamical bifurcation switches the encoding mode from redundant to synergistic, and dissipation can be used to tune between them. The empirical core is strengthened by consistency across three levels of approximation (mean-field, second-order cumulant, full quantum) and by the demonstration that the synergy peak persists for two different input statistics. The Keldysh calculation and the pedagogical PID appendices are useful and mostly clearly presented. The main weakness is that the proposed causal mechanism is computed around the undriven, empty-cavity fixed point, whereas the simulations operate at finite drive amplitudes; the paper does not yet establish that the pole structure it computes governs the driven reservoir's encoding dynamics. The memory-capacity discussion also contains claims that are only partially supported by the presented data.

major comments (3)
  1. [Section 3.2, Eqs. (12)-(13); Appendix C, Eq. (C.7)] The overdamped-soft-mode mechanism is derived by linearizing around the undriven fixed point α1,c=α2,c=0, where all Kerr terms vanish in Eq. (C.4). The simulations that exhibit the synergy peak use F=0.2 (quantum regime) or F=2.0 (mean-field regime), so the actual driven steady state is displaced. For the quantum parameters at J=|Δ|, the mean-field steady-state occupation satisfies n≈F²/((Un)²+γ²), giving n≈0.095 and a Kerr-induced frequency shift U1 n≈0.38, which is comparable to γ=0.5. This moves the would-be zero-frequency slow pole away from the imaginary axis, so Eq. (12) does not describe the linear response in the regime where the synergy peak is reported. The empirical peak may survive, but the causal claim that soft-mode overdamping drives the synergy enhancement needs to be re-derived around the driven fixed point or supported by additional evidence.
  2. [Section 3.2, final paragraph] The coherence-driven synergy enhancement is justified in the regime γ∼Re(ω_f), but the numerical quantum regime uses γ=0.5 and Re(ω_f)=|J+|Δ||=4 at J=|Δ|, so γ is not comparable to the fast-mode frequency. The mean-field regime also uses γ=0.5 with the same fast-mode frequency. The paper should either demonstrate that the mechanism operates outside the stated γ∼Re(ω_f) regime or provide simulations in a parameter regime that actually satisfies this condition.
  3. [Section 3.4, Figs. 8-9] The abstract's memory claim ('synergy ... enhances immediate memory retention, whereas strong dissipation ... supports long-term memory retention') is only partially supported by the data. Fig. 8 shows that approaching J=|Δ| improves short-delay capacity and reduces long-delay capacity, but Fig. 9 shows that at J=|Δ| increasing γ increases memory capacity at all delays, with the decay exponent Γ stated to remain approximately constant. Since long-delay memory also improves with γ in Fig. 9, the data do not demonstrate that redundant encoding preferentially supports long-term retention. The text itself concedes that the low-dissipation system has not fully achieved the fading-memory regime. The memory conclusion should be reframed to match Fig. 9, or a parameter regime should be identified where a crossover in MC(n) with γ is evident.
minor comments (5)
  1. [Section 3.4, text near Fig. 9] The sentence 'as shown in Fig. 8 while higher γ leads to improved total memory capacity' appears to refer to Fig. 9, not Fig. 8; please correct the reference.
  2. [Appendix C, Eq. (C.1)] The Keldysh action in Eq. (C.1) contains drive terms with a factor √2, whereas the mean-field equation (4) and Eq. (C.2) contain F(t) without this factor; please clarify the convention used in the Keldysh rotation.
  3. [Section 3.1, Fig. 3 and Appendix D] The three curves in Fig. 3 are computed at different parameter sets (mean-field F=2, U1=6.25×10⁻³; second-order cumulant F=0.5, U1=0.2; quantum F=0.2, U1=4). The comparison would be more convincing if the approximation schemes were compared at a common parameter set, or if the text explicitly stated why different parameters are needed.
  4. [Appendix C, final paragraph] There is a typo: 'disapperance' should be 'disappearance'.
  5. [Figure 3] The overlapping markers in Fig. 3 make it difficult to distinguish the three curves, especially near the peak; larger markers or separated panels would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the synergy peak is measured from simulated readouts and the Keldysh linear-response mechanism is an independent calculation, so the central claims do not reduce to their inputs.

full rationale

The paper's central empirical result, the redundant-to-synergistic transition near J = |Δ|, is obtained by computing PID components from simulated output readouts Xi(t) = Re⟨âi(t)⟩, which is a direct measurement of the model dynamics rather than a fit. The linear-response explanation in Section 3.2 uses the Keldysh retarded Green's function poles derived in Appendix C, Eq. (C.7), from the quadratic fluctuation action; this is a parameter-free calculation with explicitly stated assumptions and is not fitted to the PID curves. No target quantity is defined in terms of another target quantity: synergy and redundancy are independent information-theoretic measures, and the pole condition Re(ωs) = 0 at J = |Δ| follows from the Hessian of the effective potential, not from the observed synergy peak. The self-citations that appear, such as [11] and [15], are contextual or limitation-related and are not load-bearing for the central derivation. The paper explicitly notes the remaining support gap that the linear-response analysis is performed around the undriven α1,c = α2,c = 0 steady state while the simulations use finite drive F, which is a correctness and scope concern about whether the overdamped-soft-mode mechanism applies at the driven fixed point, but it is not circularity: the mechanism is an independent theoretical claim, not an input that is relabeled as a prediction. Overall, the derivation chain is self-contained against external benchmarks, and no circular step is exhibited.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

All listed parameters are hand-chosen model settings rather than quantities fit to data, and the central result depends on the regime they define. No new physical entities are introduced; the soft and fast modes are normal modes of the linearized system, not new postulated degrees of freedom.

free parameters (5)
  • Kerr nonlinearities U1, U2 = 2 U1 = Quantum regime U1=4.0, U2=8.0; mean-field regime U1=6.25e-3, U2=1.25e-2
    Unequal Kerr coefficients are introduced to break symmetry between the oscillators (Section 2.1); the specific values define the two working regimes and are chosen by hand, not fitted to external data.
  • Detuning Delta = -2
    The detuning sets the critical coupling J_c=|Delta|=2 inside the J sweep from 0 to 4; chosen by hand, not fitted.
  • Drive amplitude F = 0.2 (quantum regime), 2.0 (mean-field regime)
    The drive is chosen to be small or comparable to other energy scales so that the reservoir's intrinsic dynamics dominate; the values are selected, not fit.
  • Photon loss rate gamma = Baseline 0.5, varied upward in the quantum regime
    Dissipation is the control parameter for the redundancy/synergy crossover; the values are hand-picked rather than inferred from data.
  • Input signal parameters = Telegraph s(t) in {-1,1}, or uniform in [-1,1], with dt=0.01
    The input statistics are testing choices; the authors check robustness with uniform uncorrelated noise, so this is not a fitted parameter.
assumptions (6)
  • domain assumption The open system is described by a Lindblad master equation with independent single-photon loss to Markovian baths.
    Equation (2); this neglects non-Markovian baths, dephasing, and multi-photon loss processes.
  • domain assumption Mean-field factorization of expectation values, <a_i a_j> approximately <a_i><a_j>, is valid in the semiclassical regime.
    Used to derive Eq. (4); valid only when quantum correlations are negligible, not in the fully quantum regime.
  • ad hoc to paper Third- and fourth-order cumulants vanish in the second-order cumulant expansion.
    Appendix D truncation; the error is uncontrolled and is expected to interpolate between mean-field and exact quantum dynamics.
  • domain assumption BROJA partial information decomposition provides the correct split of total mutual information into redundancy and synergy.
    PID is not unique; the BROJA-2PID convention in Section 2.3 and Appendix B fixes one decomposition, and results may depend on that choice.
  • ad hoc to paper Linear response around the undriven steady state alpha1,c=alpha2,c=0 captures the driven reservoir's encoding dynamics.
    Appendix C, Eq. (C.7); the same alpha=0 reference is used even for F=2.0 mean-field runs, where the drive is not a small perturbation.
  • domain assumption The readout observables Xi(t)=Re<a_i(t)> and Yi(t)=Im<a_i(t)> capture the information available to a reservoir computer.
    Equation (5); other observables such as photon numbers or different quadratures could yield different PID decompositions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dissipation alters modes of information encoding in small quantum reservoirs near criticality." pith.science (2026). https://pith.science/paper/GRLT4NEH

@misc{pith2026241218290,
  author       = {Pith},
  title        = {Pith review of: Dissipation alters modes of information encoding in small quantum reservoirs near criticality},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GRLT4NEH}},
  note         = {Machine review of arXiv:2412.18290}
}
read the original abstract

Quantum reservoir computing (QRC) has emerged as a promising paradigm for harnessing near-term quantum devices to tackle temporal machine learning tasks. Yet identifying the mechanisms that underlie enhanced performance remains challenging, particularly in many-body open systems where nonlinear interactions and dissipation intertwine in complex ways. Here, we investigate a minimal model of a driven-dissipative quantum reservoir described by two coupled Kerr-nonlinear oscillators, an experimentally realizable platform that features controllable coupling, intrinsic nonlinearity, and tunable photon loss. Using Partial Information Decomposition (PID), we examine how different dynamical regimes encode input drive signals in terms of redundancy (information shared by each oscillator) and synergy (information accessible only through their joint observation). Our key results show that, near a critical point marking a dynamical bifurcation, the system transitions from predominantly redundant to synergistic encoding. We further demonstrate that synergy amplifies short-term responsiveness, thereby enhancing immediate memory retention, whereas strong dissipation leads to more redundant encoding that supports long-term memory retention. These findings elucidate how the interplay of instability and dissipation shapes information processing in small quantum systems, providing a fine-grained, information-theoretic perspective for analyzing and designing QRC platforms.

Figures

Figures reproduced from arXiv: 2412.18290 by the authors.

Figure 1
Figure 1. Schematic of two coupled Kerr-nonlinear oscillators. Each cavity [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Classical mutual information I [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Normalized synergy (left) and normalized redundancy (right) vs. the coupling [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Effective potential around the steady state [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Retarded Green’s function poles (C.7) in the complex-frequency plane as J increases, illustrating the evolution of slow modes ωs (orange dots) and fast modes ωf (blue dots). For J < |∆|, both slow and fast modes coexist, and the system tends to encode inputs more redun…
Figure 6
Figure 6. Figure 6: Impact of uniform, uncorrelated noise input on information encoding at the [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: Impact of increasing γ on information metrics in the quantum dynamics regime (∆ = −2 and F = 0.2). (a) Time-averaged quantum mutual information (QMI) between the two oscillators as a function of J. (b) Classical mutual information (MI) between the input signal and the …
Figure 8
Figure 8. Figure 8: Memory capacity as a function of n as J approaches the critical coupling at J = |∆| for the parameters ∆ = −2, γ = 0.5, and F = 0.2. (a) n = 1−10 for J ∈ [0, 2], showing an increase in short-term memory capacity as J → |∆|. (b) n = 1−20 for J ∈ [1.96, 2], showing the l…
Figure 9
Figure 9. Figure 9: Memory capacity at the critical coupling [PITH_FULL_IMAGE:figures/full_fig_p017_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 51 canonical work pages

  1. [1]

    Jaeger H 2001 GMD Report

    work page 2001

  2. [2]

    Lukoˇ seviˇ cius M and Jaeger H 2009Computer science review 3 127–149

    work page

  3. [3]

    Nakajima K 2018 IEICE Technical Report; IEICE Tech. Rep. 118 149–154

    work page 2018

  4. [4]

    Lukoˇ seviˇ cius M 2012 A practical guide to applying echo state networksNeural Networks: Tricks of the Trade: Second Edition (Springer) pp 659–686

    work page 2012

  5. [5]

    Fujii K and Nakajima K 2017 Physical Review Applied 8 024030

    work page 2017

  6. [6]

    Innocenti L, Lorenzo S, Palmisano I, Ferraro A, Paternostro M and Palma G M 2023 Communications Physics 6 118 ISSN 2399-3650 URL https://doi.org/10.1038/ s42005-023-01233-w

    work page 2023

  7. [7]

    Motamedi A, Zadeh-Haghighi H and Simon C 2023 arXiv preprint arXiv:2304.03462

    arXiv 2023

  8. [8]

    G¨ otting N, Lohof F and Gies C 2023 Physical Review A 108 052427

    work page 2023

Show all 67 references
  1. [9]

    Mart ´ ınez-Pe˜ na R and Ortega J P 2023Phys. Rev. E 107(3) 035306 URL https://link.aps. org/doi/10.1103/PhysRevE.107.035306

    doi

  2. [10]

    Sannia A, Mart ´ ınez-Pe˜ na R, Soriano M C, Giorgi G L and Zambrini R 2024 Quantum 8 1291 ISSN 2521-327X URL https://doi.org/10.22331/q-2024-03-20-1291

    2024 doi

  3. [11]

    Xiong W, Facelli G, Sahebi M, Agnel O, Chotibut T, Thanasilp S and Holmes Z 2024 On fundamental aspects of quantum extreme learning machines ( Preprint 2312.15124) URL https://arxiv.org/abs/2312.15124

    2024 arXiv

  4. [12]

    Mujal P, Mart ´ ınez-Pe˜ na R, Giorgi G L, Soriano M C and Zambrini R 2023 npj Quantum Information 9 16 ISSN 2056-6387 URL https://doi.org/10.1038/s41534-023-00682-z

    2023 doi

  5. [13]

    Govia L, Ribeill G, Rowlands G, Krovi H and Ohki T 2021 Physical Review Research 3 013077

    2021

  6. [14]

    Ivaki M N, Lazarides A and Ala-Nissila T 2024 Quantum reservoir computing on random regular graphs (Preprint 2409.03665) URL https://arxiv.org/abs/2409.03665

    2024 arXiv

  7. [15]

    Sornsaeng A, Dangniam N and Chotibut T 2024 Quantum Machine Intelligence 6 57 ISSN 2524- 4914 URL https://doi.org/10.1007/s42484-024-00188-7

    2024 doi

  8. [16]

    Mart ´ ınez-Pe˜ na R, Giorgi G L, Nokkala J, Soriano M C and Zambrini R 2021 Physical Review Letters 127 100502

    2021

  9. [17]

    wiley.com/doi/pdf/10.1002/qute.202100027) URL https://onlinelibrary.wiley.com/ doi/abs/10.1002/qute.202100027

    Mujal P, Mart ´ ınez-Pe˜ na R, Nokkala J, Garc ´ ıa-Beni J, Giorgi G L, Soriano M C and Zambrini R 2021 Advanced Quantum Technologies 4 2100027 ( Preprint https://onlinelibrary. wiley.com/doi/pdf/10.1002/qute.202100027) URL https://onlinelibrary.wiley.com/ doi/abs/10.1002/qute...

    2021 doi

  10. [18]

    Hartmann M, Brand˜ ao F and Plenio M 2008 Laser & Photonics Reviews 2 527– 556 (Preprint https://onlinelibrary.wiley.com/doi/pdf/10.1002/lpor.200810046) URL https://onlinelibrary.wiley.com/doi/abs/10.1002/lpor.200810046

    2008 doi

  11. [19]

    Gat I and Tishby N 1998 Advances in neural information processing systems 11

    1998

  12. [20]

    Brenner N, Strong S P, Koberle R, Bialek W and Steveninck R R d R v 2000 Neural computation 12 1531–1552

    2000

  13. [21]

    Schneidman E, Bialek W and Berry M J 2003 Journal of Neuroscience 23 11539–11553

    2003

  14. [22]

    Latham P E and Nirenberg S 2005 Journal of Neuroscience 25 5195–5206

    2005

  15. [23]

    Bertschinger N, Rauh J, Olbrich E, Jost J and Ay N 2014 Entropy 16 2161–2183

    2014

  16. [24]

    Kamenev A 2023 Field theory of non-equilibrium systems (Cambridge University Press)

    2023

  17. [25]

    Dudas J, Carles B, Plouet E, Mizrahi F A, Grollier J and Markovi´ c D 2023 npj Quantum Information 9 64 ISSN 2056-6387 URL https://doi.org/10.1038/s41534-023-00734-4 Dissipation alters modes of information encoding in small quantum reservoirs 31

    2023 doi

  18. [26]

    Dudas J, Grollier J and Markovi´ c D 2022 Coherently coupled quantum oscillators for quantum reservoir computing 2022 IEEE 22nd International Conference on Nanotechnology (NANO) pp 397–400

    2022

  19. [27]

    Kalfus W D, Ribeill G J, Rowlands G E, Krovi H K, Ohki T A and Govia L C G 2022 Phys. Rev. Res. 4(3) 033007 URL https://link.aps.org/doi/10.1103/PhysRevResearch.4.033007

    2022 doi

  20. [28]

    Khan S A, Hu F, Angelatos G and T¨ ureci H E 2021 Physical reservoir computing using finitely- sampled quantum systems ( Preprint 2110.13849) URL https://arxiv.org/abs/2110.13849

    2021 arXiv

  21. [29]

    Milburn G and Holmes C 1991 Physical Review A 44 4704

    1991

  22. [30]

    Milburn G, Corney J, Wright E M and Walls D 1997 Physical Review A 55 4318

    1997

  23. [31]

    Carusotto I and Ciuti C 2013 Reviews of Modern Physics 85 299

    2013

  24. [32]

    Goto H and Kanao T 2021 Physical Review Research 3 043196

    2021

  25. [33]

    2002 The theory of open quantum systems (Oxford University Press on Demand)

    Breuer H P, Petruccione F et al. 2002 The theory of open quantum systems (Oxford University Press on Demand)

    2002

  26. [34]

    Bena I 2006 International Journal of Modern Physics B 20 2825–2888

    2006

  27. [35]

    Abel B and Marquardt F 2008 Physical Review B 78 201302

    2008

  28. [36]

    Franco R L, D’Arrigo A, Falci G, Compagno G and Paladino E 2012 Physica Scripta 2012 014019

    2012

  29. [37]

    Makkeh A, Theis D O and Vicente R 2018 Entropy 20 271

    2018

  30. [38]

    Soriente M, Chitra R and Zilberberg O 2020 Physical Review A 101 023823

    2020

  31. [39]

    Soriente M, Heugel T L, Omiya K, Chitra R and Zilberberg O 2021 Physical Review Research 3 023100

    2021

  32. [40]

    Alaeian H, Soriente M, Najafi K and Yelin S 2021 arXiv preprint arXiv:2106.04045

    2021 arXiv

  33. [41]

    Dambre J, Verstraeten D, Schrauwen B and Massar S 2012Scientific Reports 2 514 ISSN 2045-2322 URL https://doi.org/10.1038/srep00514

    doi

  34. [42]

    Mart ´ ınez-Pe˜ na R and Ortega J P 2024 Input-dependence in quantum reservoir computing (Preprint 2412.08322) URL https://arxiv.org/abs/2412.08322

    2024 arXiv

  35. [43]

    Verstraete F, Wolf M M and Ignacio Cirac J 2009 Nature physics 5 633–636

    2009

  36. [44]

    Fry D, Deshmukh A, Chen S Y C, Rastunkov V and Markov V 2023 Scientific Reports 13 19326 ISSN 2045-2322 URL https://doi.org/10.1038/s41598-023-45015-4

    2023 doi

  37. [45]

    2015 Science 347 853–857

    Leghtas Z, Touzard S, Pop I M, Kou A, Vlastakis B, Petrenko A, Sliwa K M, Narla A, Shankar S, Hatridge M J et al. 2015 Science 347 853–857

    2015

  38. [46]

    Shannon C E 1948 The Bell system technical journal 27 379–423

    1948

  39. [47]

    Bell A J 2003 The co-information lattice Proceedings of the Fifth International Workshop on Independent Component Analysis and Blind Signal Separation: ICA vol 2003

    2003

  40. [48]

    James R G, Ellison C J and Crutchfield J P 2011 Chaos: An Interdisciplinary Journal of Nonlinear Science 21 037109

    2011

  41. [49]

    Harder M, Salge C and Polani D 2013 Physical Review E 87 012130

    2013

  42. [50]

    Olbrich E, Bertschinger N and Rauh J 2015 Entropy 17 3501–3517

    2015

  43. [51]

    Rosas F E, Mediano P A, Gastpar M and Jensen H J 2019 Physical Review E 100 032305

    2019

  44. [52]

    Gutknecht A J, Makkeh A and Wibral M 2023 arXiv preprint arXiv:2306.00734

    2023 arXiv

  45. [53]

    Williams P L and Beer R D 2010 arXiv preprint arXiv:1004.2515

    2010 arXiv

  46. [54]

    Bertschinger N, Rauh J, Olbrich E and Jost J 2013 Shared information—new insights and problems in decomposing information in complex systems Proceedings of the European conference on complex systems 2012 (Springer) pp 251–269

    2013

  47. [55]

    Banerjee P K and Griffith V 2015 arXiv preprint arXiv:1509.03706

    2015 arXiv

  48. [56]

    Wibral M, Priesemann V, Kay J W, Lizier J T and Phillips W A 2017 Brain and cognition 112 25–38

    2017

  49. [57]

    Makkeh A, Gutknecht A J and Wibral M 2021 Physical Review E 103 032149

    2021

  50. [58]

    Schick-Poland K, Makkeh A, Gutknecht A J, Wollstadt P, Sturm A and Wibral M 2021 arXiv preprint arXiv:2106.12393

    2021 arXiv

  51. [59]

    Rauh J, Banerjee P K, Olbrich E, Jost J and Bertschinger N 2017 Entropy 19 328

    2017

  52. [60]

    Griffith V and Koch C 2014 Guided self-organization: inception 159–190 Dissipation alters modes of information encoding in small quantum reservoirs 32

    2014

  53. [61]

    McGill W 1954 Transactions of the IRE Professional Group on Information Theory 4 93–111

    1954

  54. [62]

    Yeung R W 1991 IEEE transactions on information theory 37 466–474

    1991

  55. [63]

    2003 Physical review letters 91 238701

    Schneidman E, Still S, Berry M J, Bialek W et al. 2003 Physical review letters 91 238701

    2003

  56. [64]

    Dalla Torre E G, Diehl S, Lukin M D, Sachdev S and Strack P 2013 Physical Review A 87 023831

    2013

  57. [65]

    Sieberer L M, Buchhold M and Diehl S 2016 Reports on Progress in Physics 79 096001

    2016

  58. [66]

    Kubo R 1962 Journal of the Physical Society of Japan 17 1100–1120

    1962

  59. [67]

    S´ anchez-Barquilla M, Silva R and Feist J 2020The Journal of chemical physics 152 034108

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.