Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Thermodynamics of Quantum Reservoir Computing

T0 review · 3 major / 6 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read This paper claims that the computational peak of quantum reservoir computing near quantum phase transitions is not an empirical accident but a spectral resonance, and that this optimal performance comes with a provable thermodynamic cost.

desk verdict Exact thermodynamic bound is sound and worth publishing; the 'analytic proof' of the critical peak is an uncontrolled approximation plus a fit. read the letter →

arxiv 2607.02157 v2 pith:2GTA5FN7 submitted 2026-07-02 quant-ph cond-mat.dis-nncond-mat.quant-gascond-mat.stat-mech

Lixiang Ding , Xingze Qiu This is my paper · ORCID
classification quant-phcond-mat.dis-nncond-mat.quant-gascond-mat.stat-mech
keywords quantumreservoircomputingHolevocapacityBogoliubov-Kubo-MorimetricspectralresonancegeneralizedLandauerboundcriticalitycoherencechaotictime-seriesprediction
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 tries to establish a first-principles thermodynamic explanation of why driven open quantum systems forecast chaotic time series best near quantum critical points. It claims that the peak arises from a spectral resonance: as the intrinsic energy gap closes, the reservoir's internal transition frequencies align with the dominant frequency of the chaotic drive, and this alignment is what maximizes the Holevo capacities. It further derives a generalized Landauer bound showing that this optimal processing carries a price: the heat dissipated to the environment must cover the reservoir's entropy change plus the quantum informational dissipation, which measures retained historical data that does not help prediction. A sympathetic reader would care because this turns an empirical observation—computation at the edge of chaos—into a predictable, engineering-relevant physical mechanism with clear energetic limits.

What carries the argument

The central object is the dynamic accumulation factor G(ΔE_jk), a filtered, windowed Fourier transform of the input autocovariance with damping from thermalization and chaotic decorrelation. It is obtained by mapping Holevo capacities onto the Bogoliubov-Kubo-Mori metric under linear response, and it acts as the spectral resonance condition: G peaks when an internal energy gap ΔE_jk equals the characteristic drive frequency ω_s. A second load-bearing piece is the quantum informational dissipation χ^d, defined as the difference between the memory Holevo capacity and the predictive Holevo capacity; it appears in the identity βW_irr = χ^d at the injection step and in the generalized Landauer bo

What would settle it

Compute the exact conditional expectation of the reservoir's filtered history given the current input (or measure it experimentally) and compare with the linear projection used in the derivation; if it deviates significantly near the critical point, the predicted peak position or height would shift. Alternatively, directly measure the input's autocovariance function: if it is not a damped cosine with a Lorentzian spectral line, the resonance denominator and the peak's dependence on drive frequency would not hold.

Watch

Extended reading notes

Core claim

The paper's central claim is that the macroscopic predictive ability of a driven open many-body quantum reservoir can be traced to a microscopic spectral matching condition. In the weak-coupling linear response regime, the memory and predictive Holevo capacities are approximated by distances on the Bogoliubov-Kubo-Mori geometric manifold. The dynamic accumulation factor acts as a frequency-selective filter: it peaks only when an internal energy gap matches the characteristic frequency of the chaotic driving signal. Near a quantum critical point the primary gap closes, sweeping transition frequencies through the drive frequency, and simultaneously the transition matrix elements grow, so capac

Load-bearing premise

The analytic proof that the critical peak is a spectral resonance rests on two approximations: that the reservoir's response to the chaotic input can be summarized by a linear projection of the current signal, and that the chaotic signal's time-correlations decay as a simple damped oscillation with a Lorentzian spectral line shape—if real chaotic drives violate either, the proof of the peak would need revision, though the Landauer-style inequality itself would survive.

Editorial extensions

If this is right

  • Engineers should tune a quantum reservoir's energy spectrum to resonate with the target signal's dominant frequency, rather than simply positioning the system near a generic phase transition.
  • Continuous temporal processing in quantum devices is thermodynamically more expensive than quasi-static erasure: the accumulated non-predictive historical information adds a mandatory heat cost.
  • The critical resonance that maximizes predictive capacity also maximizes informational dissipation and the irreversible work needed for environmental erasure, so peak performance and peak dissipation come together.
  • Quantum coherences can improve prediction without extra mechanical work, at least when the driving field is diagonal in the readout basis, giving a resource-efficient quantum advantage.
  • The same spectral mechanism governs multi-step memory and multi-horizon forecasting, so the critical-region peak persists across delays and prediction horizons.

Reading between the lines

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

  • Beyond the paper: the resonance mechanism could be tested directly in existing hardware by sweeping a reservoir's energy gap at fixed drive frequency and checking whether the predictive capacity peaks exactly at ΔE ≈ ω_s.
  • Beyond the paper: the linear-projection approximation suggests that inputs with strongly non-Gaussian statistics or heavy tails may shift the optimal gap away from ω_s; a nonlinear version of the spectral filter would be a natural extension.
  • Beyond the paper: the generalized Landauer bound points toward a quantum analogue of the information bottleneck principle, where a reservoir should compress non-predictive history while preserving coherent predictive features.
  • Beyond the paper: the critical sensitivity of informational dissipation could be used as a thermodynamic witness of quantum phase transitions, turning machine-learning metrics into probes of many-body physics.
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 / 6 minor

Summary. The manuscript develops a nonequilibrium thermodynamic framework for quantum reservoir computing. It defines memory (χ_m) and predictive (χ_p) Holevo capacities conditioned on current inputs and future targets, introduces the quantum informational dissipation χ_d = χ_m − χ_p, and decomposes both capacities into classical mutual information and ensemble coherence. Using stochastic unravelings and non-equilibrium free energies, it derives Eq. (13), βW_irr = χ_d, and a generalized Landauer bound βQ_diss ≥ ΔS_sys + χ_d_tot. It then uses a BKM-metric linear-response expansion to obtain an approximate analytical expression for the capacities [Eq. (16)] in which a dynamical accumulation factor G(ΔE_jk) acts as a frequency-selective filter. The paper's central claim is that the computational peak in the quantum critical region is analytically proven to originate from a spectral resonance: as the energy gap closes, reservoir transition frequencies align with the dominant frequency of the chaotic Mackey-Glass drive. Numerical simulations on a disordered transverse-field Ising model and an augmented cluster model show capacity peaks, NMSE minima, and satisfaction of the Landauer bound, together with a coherence decomposition indicating negative coherent dissipation in the critical region.

Significance. If the central claim held, the paper would convert the empirically observed 'edge of chaos' advantage of QRC into a design principle: match the reservoir spectrum to the data's characteristic frequency, at a thermodynamic price quantified by QID. The exact parts of the paper are genuine strengths: the Landauer-type inequality follows rigorously from the stated definitions plus CPTP monotonicity of quantum relative entropy, and the coherence decomposition provides a clean, fair baseline for quantum advantage. The manuscript also provides detailed appendices and promises open code and data. However, the proof of the critical-peak mechanism is not yet at the claimed level. It rests on an uncontrolled linear-projection approximation and a fitted Lorentzian autocovariance model, and the analytic accumulation factor G(ΔE) is never quantitatively validated against the exact binned G or against the full capacity. In addition, the generalized Landauer bound is in large part an exact bookkeeping identity: Eq. (13) follows from the chosen definition of W_irr after cancellation of energy terms, so the physical content is the CPTP monotonicity bound on the relaxation contribution. This is a str

major comments (3)
  1. [§VII, Appendix C2, Eq. (C35)] The 'analytic proof' of the critical-region spectral resonance depends on replacing the nonlinear conditional expectation E_{S^{n-1}}[Y_jk(n)|s_n] by the linear projection (E[Y s]/σ_s²) s_n. This is exact only for jointly Gaussian variables or when the conditional mean is known to be linear; the Mackey-Glass drive is neither. The text itself acknowledges at the start of Appendix C2 that the exact conditional expectation is 'generally intractable.' Yet no validation is provided. Figure 3(a) reports a binned G(ω), but the manuscript never compares that exact binned quantity with the analytic G(ΔE) from Eqs. (C39)–(C42). Without such a comparison, the calculation supports, but does not prove, the resonance mechanism. Please add a direct numerical comparison of the exact binned G(ΔE_jk) with the approximate analytic form across the relevant transition gaps, and quantify the approximation err
  2. [Appendix C2, Eq. (C40)] The Lorentzian autocovariance model C_a(τ)=σ_s² cos(ω_s τδt)e^{-γ_s τδt} is an assumed form, not derived from the Mackey-Glass dynamics. The parameters ω_s and γ_s are read off from the numerical spectrum, and the later peak value G(ω_s)≈σ_s²/[4(P_th+γ_sδt)^2] depends directly on this model. A chaotic spectrum is broadband and not guaranteed to have a Lorentzian line shape; the autocovariance could deviate significantly in the tail that matters for the geometric series in Eq. (C41). This is a load-bearing phenomenological input for the central claim. Please validate the model by direct autocovariance estimation and by showing that the resonance prediction is robust to deviations from the Lorentzian form; otherwise the 'prediction' is partly a fit to the data used for validation.
  3. [§VII, Fig. 2 and Fig. 3] The analytical capacity formula, Eq. (16), is never compared with the exact numerical capacities plotted in Fig. 2. The paper also does not show that, at the critical parameters J≈2.5 and α≈0.5, there actually exist low-lying transition gaps with ΔE_jk≈ω_s carrying significant weight in Eq. (16). The spectral plots in Fig. 3(b,c) are qualitative; they are not overlaid with G(ω) or weighted by the transition matrix elements. To substantiate the claim that the computational peak is driven by spectral resonance rather than solely by the growth of |⟨j|H1|k⟩|² or other critical effects, please provide a quantitative decomposition of the exact capacity into spectral windows and compare it with the analytic prediction.
minor comments (6)
  1. [Fig. 3(a) caption] The caption should state explicitly whether the plotted G(ω) is the exact binned conditional variance, the analytic Lorentzian formula, or the Fourier power spectrum. Currently 'G(ω) calculated using probabilities estimated via the binning method' is ambiguous.
  2. [Appendix C2, after Eq. (C40)] Report the fitted values and the fitting procedure for ω_s and γ_s. The values γ_sδt≈0.015 and σ_s²≈0.11 appear later in the text but are not tied to a clearly described estimation method.
  3. [Abstract and §VII] The phrase 'analytically prove' in the abstract and Sec. VII is stronger than what the derivation supports, given the stated approximations in Appendix C2. Consider softening to 'analytically show' or adding the qualifications in the abstract.
  4. [Reference [75]] The data/code availability reference is listed as a citation without a URL or DOI. Please provide a working identifier or repository link.
  5. [Section V] The term 'quantum informational dissipation' could be misread as physical entropy production. A sentence clarifying that χ_d is an information-theoretic difference that is connected to heat only through Eq. (13) would improve readability.
  6. [Fig. 2] Given that averages are taken over 5000 sequences and 100 disorder realizations, standard errors or shaded confidence bands would strengthen the numerical claims, especially near the critical peaks.

Circularity Check

2 steps flagged · score 6.0 of 10

Critical-peak 'analytic proof' is partly a fit: ω_s and γ_s are read from the MG spectrum and then used to compute G(ω_s); the generalized Landauer bound is a definitional identity.

  1. fitted input called prediction [Appendix C2, Eqs. (C40)-(C42); Section VII]
    "Modeling this broadened spectral peak using a Lorentzian line shape centered symmetrically at ±ω_s, the inverse discrete-time Fourier transform directly yields an exponentially decaying harmonic oscillation: C_a(τ)≈σ_s^2 cos(ω_sτδt)e^{-γ_sτδt}. ... For our selected parameters (P_th≈0.095, γ_sδt≈0.015, and σ_s^2≈0.11), this theoretical maximum evaluates to G(ω_s)≈2.3, consistent with the empirical peak shown in Fig. 3(a)."

    The 'theoretical maximum' of the spectral-accumulation factor is evaluated from a Lorentzian autocovariance whose parameters ω_s, γ_s, and σ_s² are themselves extracted from the same MG numerical spectrum (ω_s≈0.36 and γ_sδt≈0.015 are quoted from Fig. 3(a) and the simulation). The agreement of G(ω_s)≈2.3 with the empirically binned peak is therefore a consistency check on a fitted model, not an independent prediction. The load-bearing conditional mean in G(ΔE)=Var(E[Y|s_n]) is also replaced by the linear projection (C35) without a comparison to the exactly binned G, so the claimed analytic proof of critical-region resonance is not established independently of these fitted/approximate inputs.

  2. self definitional [Section V, Eq. (13); Appendix F, Eq. (F3)]
    "βW^{irr}_{n+1}=Σ_{s_{n+1}}P(s_{n+1})S(ρ^p_{s_{n+1}})−Σ_{s_n}P(s_n)S(ρ^m_{s_n}) (Appendix F). Recalling Eq. (10), this entropic cost translates directly to the QID: βW^{irr}_{n+1}=χ^d_{t_{n+1}}. (13)"

    W^{irr} is defined as W−ΔF, and ΔF is constructed from the same predictive/memory conditional entropies; substituting the definitions cancels all energy terms and leaves exactly χ^d=χ^m−χ^p. Consequently Eqs. (14)-(15) are algebraic rearrangements of definitions plus the non-negativity of the relaxation dissipation (CPTP monotonicity), rather than an independently predicted physical law. The bound cannot fail for any reservoir because the equality at each step is fixed by how W^{irr} and χ^d are defined.

full rationale

The exact parts of the paper—Holevo capacity definitions, coherence decomposition (χ=I+C), and the CPTP-monotonicity argument giving βW_relax≥0—are internally consistent mathematical consequences and do not reduce to fitted inputs. There is no load-bearing self-citation: [75] is only a code/data link, and the cited theorems (BKM metric, CPTP monotonicity, linear projection) are standard external results. The central headline claim, however, is the 'analytic proof' of the spectral-resonance peak. That proof passes through Eq. (C40), a Lorentzian autocovariance whose ω_s and γ_s are taken from the numerical MG spectrum, and Eq. (C42), which uses those fitted values to obtain G(ω_s)≈2.3 and calls agreement with the empirical peak a confirmation. The exact conditional mean in G is also replaced by a linear projection (C35) that is only guaranteed for Gaussian statistics; the paper does not compare this approximate G with the exact binned G shown in Fig. 3(a). Thus the quantitative 'prediction' is partly a refit of the data it claims to explain, while the qualitative resonance mechanism remains plausible but not independently proven. The Landauer inequality itself is a definitional identity (βW_irr=χ^d) plus non-negativity, so its claim to be a new fundamental bound is more a formal reformulation than an independent law. Overall, partial circularity in the central critical-peak prediction: score 6.

Assumptions & free parameters 4 free parameters · 6 assumptions · 1 invented entities

The exact thermodynamic parts rest on standard quantum-information identities plus the collision-model thermalization assumption. The critical-peak explanation additionally imports several modeling/fitting choices: the linear-projection approximation, the Lorentzian autocovariance, and the data-estimated ω_s and γ_s. These are load-bearing for the headline 'analytic proof' but not for the generalized Landauer bound itself.

free parameters (4)
  • coupling strength λ = 0.05
    Chosen small to justify linear response; not fitted to a target but enters all capacity formulas quadratically.
  • thermalization probability P_th = 1−exp(−0.1) ≈ 0.095 per step
    Set by γ0=0.1 and δt=1; controls memory fading and appears directly in the G(ω_s) peak prediction.
  • MG characteristic frequency ω_s = ≈0.36
    Estimated from the numerical power spectrum of the Mackey-Glass sequence; the resonance condition ΔE≈ω_s is the central explanatory mechanism.
  • chaotic decorrelation rate γ_s = ≈0.015/δt
    Obtained by fitting a Lorentzian line shape to the MG autocovariance; used together with P_th to predict G(ω_s)≈2.3.
assumptions (6)
  • standard math Holevo quantity, quantum relative entropy identities, and CPTP monotonicity of relative entropy
    Invoked in Secs. III and V; standard results from quantum information theory.
  • domain assumption Discrete thermalization map ρ→(1−P_th)UρU†+P_thρ_eq (Eq. A1) is a faithful model of the open-system dynamics
    A collision-model choice, not derived from a specific physical bath; the global thermodynamic consistency is inherited from the secular global master equation.
  • ad hoc to paper Linear projection of the conditional expectation E_{S^{n-1}}[Y_{jk}(n)|s_n] onto c·s_n (Eq. C35)
    For a nonlinear chaotic input the conditional mean need not be linear; no Gaussianity or linearity justification is given. This underpins Eq. (C39).
  • ad hoc to paper Mackey-Glass autocovariance has the damped-harmonic Lorentzian form C_a(τ)=σ_s² cos(ω_s τδt)e^{−γ_sτδt} (Eq. C40)
    A model assumption for the chaotic power spectrum; ω_s and γ_s are estimated from the data, making the later G(ω_s) agreement partly a fit.
  • domain assumption The chaotic input sequence is zero-mean and stationary
    Verified empirically in Fig. 3(a2); used to derive the linear-response baseline and the variance interpretation of G.
  • domain assumption Weak-coupling linear response is valid at λ=0.05, including near criticality
    The BKM quadratic approximation assumes small perturbations; critical systems can have enhanced susceptibility, so the linear-response truncation is not automatically safe.
invented entities (1)
  • quantum informational dissipation χ_d = χ^m − χ^p independent evidence
    purpose: Metric for the non-predictive historical information retained by the reservoir; appears in Eq. (10) and in the generalized Landauer bound Eq. (15).
    Directly computable from conditional state ensembles and testable by tomography, but it is a derived information-theoretic quantity rather than a new physical object.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Thermodynamics of Quantum Reservoir Computing." pith.science (2026). https://pith.science/paper/2GTA5FN7

@misc{pith2026260702157,
  author       = {Pith},
  title        = {Pith review of: Thermodynamics of Quantum Reservoir Computing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2GTA5FN7}},
  note         = {Machine review of arXiv:2607.02157}
}
read the original abstract

Quantum reservoir computing provides a framework for processing complex temporal data, yet its fundamental computational and energetic limits remain unresolved. Here, we establish a non-equilibrium thermodynamic framework that links the macroscopic predictive performance of driven open quantum systems to their microscopic energetic costs. By mapping Holevo capacities onto the Bogoliubov-Kubo-Mori geometric manifold, we analytically prove that the computational peak within the quantum critical region originates from a spectral resonance: the closing of the intrinsic energy gap forces the reservoir's internal transition frequencies to align with the chaotic drive. To evaluate the associated thermodynamic costs, we introduce quantum informational dissipation to quantify the non-predictive historical data retained by the reservoir. This allows us to derive a generalized Landauer bound for continuous temporal processing, which reveals a fundamental thermodynamic trade-off: the critical resonance that maximizes predictive capacity simultaneously maximizes informational dissipation and the irreversible work required for environmental erasure. Furthermore, coherence decomposition demonstrates that quantum coherences amplify predictive capacity without demanding additional mechanical work. These findings establish the fundamental energetic limits of quantum learning devices, providing theoretical principles for designing energy-efficient quantum neuromorphic hardware.

Figures

Figures reproduced from arXiv: 2607.02157 by the authors.

Figure 1
Figure 1. FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Storage, Scrambling, and Loss of Information in the Quantum Reservoir Computing Paradigm

    quant-ph 2026-08 conditional novelty 6.0 of 10

    Quantum reservoir computing can be characterized by a classical-quantum state whose Holevo quantities yield effective scrambling and memory-decay diagnostics that track the memory-nonlinearity trade-off in an Ising reservoir.

Reference graph

Works this paper leans on

88 extracted references · 1 linked inside Pith · cited by 1 Pith paper

  1. [1]

    X Sn P(S n)sm # (1−P th)n−mU n−m(V) = nX m=0

    Analytical capacities for the discrete thermalization map To evaluate the macroscopic informational capacities while preserving the non-Markovian complexity of the classical driving sequence, we perform a direct perturbative expansion on the discrete thermalization map in the weak-coupling limit (λ→0). We first establish the statistical consistency of the...

    2000

  2. [2]

    ∆Sc sys + N−1X n=0 Dc tn+1 # | {z } Lc +

    Microscopic origin of the critical computational peak Empirical observations across various quantum reservoir architectures indicate that the macroscopic information capacities (χ m tn+1 andχ p tn+1 ) simultaneously maximize within the quantum critical region. To identify the physical mechanisms driving this phenomenon, we analyze the microscopic analytic...

  3. [3]

    Throughout this work, we setβ MG = 0.2,γ MG = 0.1, andτ MG = 18 to ensure chaotic dynamics [9]

    Chaotic time-series generation The continuous classical input sequence is generated using the MG delay differential equation [70]: ds(t) dt = βMG s(t−τ MG) 1 +s 10(t−τ MG) −γ MG s(t).(G1) Originally introduced to model nonlinear physiological feedback systems, the MG equation has become a standard benchmark for time-series prediction tasks. Throughout thi...

  4. [4]

    (8)] as an example

    Estimation of Holevo capacities We outline the numerical procedure taking the quantum memory capacityχ m tn+1 =S(¯ρtn+1 )− P sn P(s n)S(ρm sn ) [Eq. (8)] as an example. The unconditional average state is ¯ρ tn+1 = P sn P(s n)ρm sn [Eq. (7)], and the conditional memory stateρ m sn aggregates the reservoir configurations over all historical trajectoriesS n−...

  5. [5]

    Cheng, A

    C. Cheng, A. Sa-Ngasoongsong, O. Beyca, T. Le, H. Yang, Z. J. Kong, and S. T. Bukkapatnam, Time series forecasting for nonlinear and non-stationary processes: a review and comparative study, IIE Transactions47, 1053 (2015)

    2015

  6. [6]

    Sch¨ otz, A

    C. Sch¨ otz, A. White, M. Gelbrecht, and N. Boers, Machine learning for predicting chaotic systems, Chaos: An Interdisci- plinary Journal of Nonlinear Science36, 053105 (2026)

    2026

  7. [7]

    echo state

    H. Jaeger,The “echo state” Approach to Analysing and Training Recurrent Neural Networks, Tech. Rep. 148 (German National Research Center for Information Technology (GMD), Bonn, Germany, 2001)

    2001

  8. [8]

    Maass, T

    W. Maass, T. Natschl¨ ager, and H. Markram, Real-Time Computing Without Stable States: A New Framework for Neural Computation Based on Perturbations, Neural Computation14, 2531 (2002)

    2002

Show all 88 references
  1. [9]

    Pathak, B

    J. Pathak, B. Hunt, M. Girvan, Z. Lu, and E. Ott, Model-Free Prediction of Large Spatiotemporally Chaotic Systems from Data: A Reservoir Computing Approach, Phys. Rev. Lett.120, 024102 (2018)

    2018

  2. [10]

    Nakajima and I

    K. Nakajima and I. Fischer,Reservoir computing: Theory, physical implementations, and applications(Springer Singapore, 2021)

    2021

  3. [11]

    M. Yan, C. Huang, P. Bienstman, P. Tino, W. Lin, and J. Sun, Emerging opportunities and challenges for the future of reservoir computing, Nature Communications15, 2056 (2024)

  4. [12]

    Zolfagharinejad, J

    M. Zolfagharinejad, J. B¨ uchel, L. Cassola, S. Kinge, G. S. Syed, A. Sebastian, and W. G. van der Wiel, Analogue speech recognition based on physical computing, Nature645, 886 (2025)

    2025

  5. [13]

    Fujii and K

    K. Fujii and K. Nakajima, Harnessing Disordered-Ensemble Quantum Dynamics for Machine Learning, Phys. Rev. Appl. 8, 024030 (2017)

    2017

  6. [14]

    Ghosh, T

    S. Ghosh, T. Paterek, and T. C. H. Liew, Quantum Neuromorphic Platform for Quantum State Preparation, Phys. Rev. Lett.123, 260404 (2019)

    2019

  7. [15]

    Ghosh, A

    S. Ghosh, A. Opala, M. Matuszewski, T. Paterek, and T. C. H. Liew, Quantum reservoir processing, npj Quantum Information5, 35 (2019)

    2019

  8. [16]

    Mujal, R

    P. Mujal, R. Mart ´ ınez-Pe˜ na, J. Nokkala, J. Garc ´ ıa-Beni, G. L. Giorgi, M. C. Soriano, and R. Zambrini, Opportunities in Quantum Reservoir Computing and Extreme Learning Machines, Advanced Quantum Technologies4, 2100027 (2021)

    2021

  9. [17]

    Mart ´ ınez-Pe˜ na, G

    R. Mart ´ ınez-Pe˜ na, G. L. Giorgi, J. Nokkala, M. C. Soriano, and R. Zambrini, Dynamical Phase Transitions in Quantum Reservoir Computing, Phys. Rev. Lett.127, 100502 (2021)

    2021

  10. [18]

    Angelatos, S

    G. Angelatos, S. A. Khan, and H. E. T¨ ureci, Reservoir Computing Approach to Quantum State Measurement, Phys. Rev. X11, 041062 (2021)

    2021

  11. [19]

    R. A. Bravo, K. Najafi, X. Gao, and S. F. Yelin, Quantum Reservoir Computing Using Arrays of Rydberg Atoms, PRX Quantum3, 030325 (2022)

    2022

  12. [20]

    Kubota, Y

    T. Kubota, Y. Suzuki, S. Kobayashi, Q. H. Tran, N. Yamamoto, and K. Nakajima, Temporal information processing induced by quantum noise, Phys. Rev. Res.5, 023057 (2023)

    2023

  13. [21]

    Dudas, B

    J. Dudas, B. Carles, E. Plouet, F. A. Mizrahi, J. Grollier, and D. Markovi´ c, Quantum reservoir computing implementation on coherently coupled quantum oscillators, npj Quantum Information9, 64 (2023)

    2023

  14. [22]

    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. 25 McMahon, Microwave signal processing using an analog quantum reservoir computer, Nature Communications15, 7490 (2024)

    2024

  15. [23]

    F. Hu, S. A. Khan, N. T. Bronn, G. Angelatos, G. E. Rowlands, G. J. Ribeill, and H. E. T¨ ureci, Overcoming the coherence time barrier in quantum machine learning on temporal data, Nature Communications15, 7491 (2024)

    2024

  16. [24]

    Ahmed, F

    O. Ahmed, F. Tennie, and L. Magri, Prediction of chaotic dynamics and extreme events: A recurrence-free quantum reservoir computing approach, Phys. Rev. Res.6, 043082 (2024)

    2024

  17. [25]

    Kobayashi, K

    K. Kobayashi, K. Fujii, and N. Yamamoto, Feedback-Driven Quantum Reservoir Computing for Time-Series Analysis, PRX Quantum5, 040325 (2024)

    2024

  18. [26]

    Y. Kora, H. Zadeh-Haghighi, T. C. Stewart, K. Heshami, and C. Simon, Frequency- and dissipation-dependent entangle- ment advantage in spin-network quantum reservoir computing, Phys. Rev. A110, 042416 (2024)

    2024

  19. [27]

    Sannia, R

    A. Sannia, R. Mart ´ ınez-Pe˜ na, M. C. Soriano, G. L. Giorgi, and R. Zambrini, Dissipation as a resource for Quantum Reservoir Computing, Quantum8, 1291 (2024)

    2024

  20. [28]

    C. Zhu, P. J. Ehlers, H. I. Nurdin, and D. Soh, Practical few-atom quantum reservoir computing, Phys. Rev. Res.7, 023290 (2025)

    2025

  21. [29]

    G¨ otting, S

    N. G¨ otting, S. Wilksen, A. Steinhoff, F. Lohof, and C. Gies, Connection between Memory Performance and Optical Absorption in Quantum Reservoir Computing, Phys. Rev. Lett.135, 240403 (2025)

    2025

  22. [30]

    Kobayashi and Y

    K. Kobayashi and Y. Motome, Quantum reservoir probing of quantum phase transitions, Nature Communications16, 3871 (2025)

    2025

  23. [31]

    D. Zia, L. Innocenti, G. Minati, S. Lorenzo, A. Suprano, R. D. Bartolo, N. Spagnolo, T. Giordani, V. Cimini, G. M. Palma, A. Ferraro, F. Sciarrino, and M. Paternostro, Quantum reservoir computing for photonic entanglement witnessing, Science Advances11, eady7987 (2025)

    2025

  24. [32]

    Kobayashi and Y

    K. Kobayashi and Y. Motome, Edge of Many-Body Quantum Chaos in Quantum Reservoir Computing, Phys. Rev. Lett. 136, 040602 (2026)

    2026

  25. [33]

    Y. Hou, J. Hua, Z. Wu, W. Xia, Y. Chen, X. Li, Z. Li, X. Peng, and J. Du, High-Accuracy Temporal Prediction via Experimental Quantum Reservoir Computing in Correlated Spins, Phys. Rev. Lett.136, 120602 (2026)

    2026

  26. [34]

    Liu, Q.-X

    D.-S. Liu, Q.-X. Jie, C.-L. Zou, X.-F. Ren, and G.-C. Guo, Practical quantum reservoir computing in Rydberg atom arrays, Phys. Rev. A113, 042401 (2026)

    2026

  27. [35]

    Di Bartolo, S

    R. Di Bartolo, S. Piacentini, F. Ceccarelli, G. Corrielli, R. Osellame, V. Cimini, and F. Sciarrino, Time-series forecasting with multiphoton quantum states and integrated photonics, npj Quantum Information12, 91 (2026)

    2026

  28. [36]

    Paparelle, J

    I. Paparelle, J. Henaff, J. Garc ´ ıa-Beni,´E. Gillet, D. Montesinos, G. L. Giorgi, M. C. Soriano, R. Zambrini, and V. Parigi, Experimental memory control in continuous-variable optical quantum reservoir computing, Nature Photonics20, 413 (2026)

    2026

  29. [37]

    Sannia, R

    A. Sannia, R. R. Rodr ´ ıguez, G. L. Giorgi, and R. Zambrini, Non-Markovianity and memory enhancement in quantum reservoir computing, npj Quantum Information 10.1038/s41534-026-01257-4 (2026)

    2026 doi

  30. [38]

    Landauer, Irreversibility and Heat Generation in the Computing Process, IBM Journal of Research and Development 5, 183 (1961)

    R. Landauer, Irreversibility and Heat Generation in the Computing Process, IBM Journal of Research and Development 5, 183 (1961)

    1961

  31. [39]

    Sagawa and M

    T. Sagawa and M. Ueda, Minimal Energy Cost for Thermodynamic Information Processing: Measurement and Information Erasure, Phys. Rev. Lett.102, 250602 (2009)

    2009

  32. [40]

    Still, D

    S. Still, D. A. Sivak, A. J. Bell, and G. E. Crooks, Thermodynamics of prediction, Phys. Rev. Lett.109, 120604 (2012)

    2012

  33. [41]

    Goold, M

    J. Goold, M. Paternostro, and K. Modi, Nonequilibrium Quantum Landauer Principle, Phys. Rev. Lett.114, 060602 (2015)

    2015

  34. [42]

    Van Vu and K

    T. Van Vu and K. Saito, Finite-Time Quantum Landauer Principle and Quantum Coherence, Phys. Rev. Lett.128, 010602 (2022)

    2022

  35. [43]

    Hsieh, Dynamical Landauer Principle: Quantifying Information Transmission by Thermodynamics, Phys

    C.-Y. Hsieh, Dynamical Landauer Principle: Quantifying Information Transmission by Thermodynamics, Phys. Rev. Lett. 134, 050404 (2025)

    2025

  36. [44]

    A. S. Holevo, Bounds for the quantity of information transmitted by a quantum communication channel, Problemy Peredachi Informatsii9, 3 (1973)

    1973

  37. [45]

    Schumacher and M

    B. Schumacher and M. D. Westmoreland, Sending classical information via noisy quantum channels, Phys. Rev. A56, 131 (1997)

    1997

  38. [46]

    Holevo, The capacity of the quantum channel with general signal states, IEEE Transactions on Information Theory44, 269 (1998)

    A. Holevo, The capacity of the quantum channel with general signal states, IEEE Transactions on Information Theory44, 269 (1998)

    1998

  39. [47]

    A. S. Holevo,Statistical structure of quantum theory, Vol. 67 (Springer Berlin, Heidelberg, 2001)

    2001

  40. [48]

    Vedral, The role of relative entropy in quantum information theory, Rev

    V. Vedral, The role of relative entropy in quantum information theory, Rev. Mod. Phys.74, 197 (2002)

    2002

  41. [49]

    M. A. Nielsen and I. L. Chuang,Quantum Computation and Quantum Information: 10th Anniversary Edition(Cambridge University Press, Cambridge, England, 2010)

    2010

  42. [50]

    M. M. Wilde,Quantum Information Theory, 2nd ed. (Cambridge University Press, 2017)

    2017

  43. [51]

    Petz, Geometry of canonical correlation on the state space of a quantum system, Journal of Mathematical Physics35, 780 (1994)

    D. Petz, Geometry of canonical correlation on the state space of a quantum system, Journal of Mathematical Physics35, 780 (1994)

    1994

  44. [52]

    Petz, Monotone metrics on matrix spaces, Linear Algebra and its Applications244, 81 (1996)

    D. Petz, Monotone metrics on matrix spaces, Linear Algebra and its Applications244, 81 (1996)

    1996

  45. [53]

    Y. G. Rudoy, Bogoliubov’s metric as a global characteristic of the family of metrics in the Hilbert algebra of observables, Theoretical and Mathematical Physics160, 1161 (2009)

    2009

  46. [54]

    Baumgratz, M

    T. Baumgratz, M. Cramer, and M. B. Plenio, Quantifying Coherence, Phys. Rev. Lett.113, 140401 (2014)

    2014

  47. [55]

    Streltsov, G

    A. Streltsov, G. Adesso, and M. B. Plenio, Colloquium: Quantum coherence as a resource, Rev. Mod. Phys.89, 041003 26 (2017)

    2017

  48. [56]

    Lecamwasam, S

    R. Lecamwasam, S. Assad, J. J. Hope, P. K. Lam, J. Thompson, and M. Gu, Relative Entropy of Coherence Quantifies Performance in Bayesian Metrology, PRX Quantum5, 030303 (2024)

    2024

  49. [57]

    Lindblad, On the generators of quantum dynamical semigroups, Communications in Mathematical Physics48, 119 (1976)

    G. Lindblad, On the generators of quantum dynamical semigroups, Communications in Mathematical Physics48, 119 (1976)

    1976

  50. [58]

    Gorini, A

    V. Gorini, A. Kossakowski, and E. C. G. Sudarshan, Completely positive dynamical semigroups ofN-level systems, Journal of Mathematical Physics17, 821 (1976)

    1976

  51. [59]

    Breuer and F

    H.-P. Breuer and F. Petruccione,The Theory of Open Quantum Systems(Oxford University Press, 2007)

    2007

  52. [60]

    Alicki, On the detailed balance condition for non-Hamiltonian systems, Reports on Mathematical Physics10, 249 (1976)

    R. Alicki, On the detailed balance condition for non-Hamiltonian systems, Reports on Mathematical Physics10, 249 (1976)

    1976

  53. [61]

    Levy and R

    A. Levy and R. Kosloff, The local approach to quantum transport may violate the second law of thermodynamics, Euro- physics Letters107, 20004 (2014)

    2014

  54. [62]

    P. P. Hofer, M. Perarnau-Llobet, L. D. M. Miranda, G. Haack, R. Silva, J. B. Brask, and N. Brunner, Markovian master equations for quantum thermal machines: local versus global approach, New Journal of Physics19, 123037 (2017)

    2017

  55. [63]

    Cattaneo, G

    M. Cattaneo, G. L. Giorgi, S. Maniscalco, and R. Zambrini, Local versus global master equation with common and separate baths: superiority of the global approach in partial secular approximation, New Journal of Physics21, 113045 (2019)

    2019

  56. [64]

    Dann and R

    R. Dann and R. Kosloff, Quantum thermo-dynamical construction for driven open quantum systems, Quantum5, 590 (2021)

    2021

  57. [65]

    T. A. Brun, Continuous measurements, quantum trajectories, and decoherent histories, Phys. Rev. A61, 042107 (2000)

    2000

  58. [66]

    Elouard, D

    C. Elouard, D. A. Herrera-Mart ´ ı, M. Clusel, and A. Auff` eves, The role of quantum measurement in stochastic thermody- namics, npj Quantum Information3, 9 (2017)

    2017

  59. [67]

    J. J. Alonso, E. Lutz, and A. Romito, Thermodynamics of Weakly Measured Quantum Systems, Phys. Rev. Lett.116, 080403 (2016)

    2016

  60. [68]

    Strasberg, Operational approach to quantum stochastic thermodynamics, Phys

    P. Strasberg, Operational approach to quantum stochastic thermodynamics, Phys. Rev. E100, 022127 (2019)

    2019

  61. [69]

    Gaveau and L

    B. Gaveau and L. Schulman, A general framework for non-equilibrium phenomena: the master equation and its formal consequences, Physics Letters A229, 347 (1997)

    1997

  62. [70]

    D. A. Sivak and G. E. Crooks, Near-Equilibrium Measurements of Nonequilibrium Free Energy, Phys. Rev. Lett.108, 150601 (2012)

    2012

  63. [71]

    J. M. R. Parrondo, J. M. Horowitz, and T. Sagawa, Thermodynamics of information, Nature Physics11, 131 (2015)

    2015

  64. [72]

    M. M. Wolf, G. Ortiz, F. Verstraete, and J. I. Cirac, Quantum Phase Transitions in Matrix Product Systems, Phys. Rev. Lett.97, 110403 (2006)

    2006

  65. [73]

    Smith, B

    A. Smith, B. Jobst, A. G. Green, and F. Pollmann, Crossing a topological phase transition with a quantum computer, Phys. Rev. Res.4, L022020 (2022)

    2022

  66. [74]

    M. C. Mackey and L. Glass, Oscillation and Chaos in Physiological Control Systems, Science197, 287 (1977)

    1977

  67. [75]

    Mascarenhas, H

    E. Mascarenhas, H. Bragan¸ ca, R. Dorner, M. Fran ¸ ca Santos, V. Vedral, K. Modi, and J. Goold, Work and quantum phase transitions: Quantum latency, Phys. Rev. E89, 062103 (2014)

    2014

  68. [76]

    A. D. Varizi, A. P. Vieira, C. Cormick, R. C. Drumond, and G. T. Landi, Quantum coherence and criticality in irreversible work, Phys. Rev. Res.2, 033279 (2020)

    2020

  69. [77]

    Tishby, F

    N. Tishby, F. C. Pereira, and W. Bialek, The information bottleneck method (2000), arXiv:physics/0004057 [physics.data- an]

    2000 arXiv

  70. [78]

    K. A. Murphy and D. S. Bassett, Machine-Learning Optimized Measurements of Chaotic Dynamical Systems via the Information Bottleneck, Phys. Rev. Lett.132, 197201 (2024)

    2024

  71. [79]

    Ding and X

    L. Ding and X. Qiu, Open-source code and data for Thermodynamics of Quantum Reservoir Computing (2026)

    2026

  72. [80]

    Scarani, M

    V. Scarani, M. Ziman, P. ˇStelmachoviˇ c, N. Gisin, and V. Buˇ zek, Thermalizing Quantum Machines: Dissipation and Entanglement, Phys. Rev. Lett.88, 097905 (2002)

    2002

  73. [81]

    Strasberg, G

    P. Strasberg, G. Schaller, T. Brandes, and M. Esposito, Quantum and Information Thermodynamics: A Unifying Frame- work Based on Repeated Interactions, Phys. Rev. X7, 021003 (2017)

    2017

  74. [82]

    Ciccarello, S

    F. Ciccarello, S. Lorenzo, V. Giovannetti, and G. M. Palma, Quantum collision models: Open system dynamics from repeated interactions, Physics Reports954, 1 (2022)

    2022

  75. [83]

    Moon and W

    T. Moon and W. Stirling,Mathematical Methods and Algorithms for Signal Processing(Prentice Hall, 2000)

    2000

  76. [84]

    Chatfield,The Analysis of Time Series: An Introduction(Chapman and Hall, 1989)

    C. Chatfield,The Analysis of Time Series: An Introduction(Chapman and Hall, 1989)

    1989

  77. [85]

    J. P. Eckmann and D. Ruelle, Ergodic theory of chaos and strange attractors, Rev. Mod. Phys.57, 617 (1985)

    1985

  78. [86]

    Kantz and T

    H. Kantz and T. Schreiber,Nonlinear Time Series Analysis, 2nd ed. (Cambridge University Press, Cambridge, 2003)

    2003

  79. [87]

    Moddemeijer, On estimation of entropy and mutual information of continuous distributions, Signal Processing16, 233 (1989)

    R. Moddemeijer, On estimation of entropy and mutual information of continuous distributions, Signal Processing16, 233 (1989)

    1989

  80. [88]

    Paninski, Estimation of Entropy and Mutual Information, Neural Computation15, 1191 (2003)

    L. Paninski, Estimation of Entropy and Mutual Information, Neural Computation15, 1191 (2003)

    2003

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.