Pith. sign in

REVIEW 4 major objections 4 minor 1 cited by

Dissipation-induced Quantum Homogenization for Temporal Information Processing

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The quantum homogenizer—an input qubit colliding one by one with identical reservoir qubits via a partial SWAP—satisfies the stability and contractivity conditions needed for quantum reservoir computing.

desk verdict A neat conceptual bridge between quantum homogenization and QRC, but the central contractivity proof doesn't go through as written, and there are no benchmarks. read the letter →

arxiv 2412.09979 v2 pith:PGGU4ZOS submitted 2024-12-13 quant-ph

Alexander Yosifov , Aditya Iyer , Vlatko Vedral This is my paper · ORCID
classification quant-ph MSC 81P6868T05 PACS 03.67.-a
keywords quantumhomogenizerreservoircomputingpartialSWAPcontractivityasymptoticstabilitydissipativedynamicstemporalinformationprocessingechostateproperty
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

Quantum reservoir computing needs a physical system whose dynamics are asymptotically stable and contractive, so that inputs are mapped into a higher-dimensional state space while memory of past inputs fades in a controlled way. This paper argues that the quantum homogenizer—an input qubit that sequentially collides with identical reservoir qubits through a time-varying partial SWAP interaction—satisfies both conditions. The authors model each collision by the completely positive trace preserving (CPTP) map $M = p\mathbb{1} + (1-p)\mathrm{SWAP}$ and claim it is contractive in the $L^2$ norm, driving every input state toward a preparable steady state. If the claim holds, the homogenizer, already realizable with NMR spin ensembles and photonic circuits, becomes a reservoir computer that needs no fine-tuned Hamiltonian, only control of interaction time.

What carries the argument

The central object is the partial SWAP unitary $\tilde{U} = \exp\!\left(-\frac{i}{\hbar}\,\hat{s}\cdot R_k \int J(t)\,dt\right)$, whose action is randomized by drawing the coupling $J(t)$ from a uniform distribution each timestep; when the integrated coupling is near $\pi$ the gate acts as a SWAP and when near zero as the identity. The associated CPTP map $M = p\mathbb{1} + (1-p)\mathrm{SWAP}$ is a convex combination, that is, a noisy quantum channel, and its repeated application defines the reservoir's discrete-time state transition. The mechanism that carries the argument is the distance inequality $\|D(k)\|_2 \le \|M(1)\|_2 \|D(k-1)\|_2$, which the paper uses to conclude that the $L^2$ distance between any two input states contracts to zero and the dynamics converge to the fixed point $\xi(0)$.

What would settle it

Numerically compute the $L^2$ operator norm of $M = p\mathbb{1} + (1-p)\mathrm{SWAP}$ restricted to the traceless symmetric subspace for $p \in (0,1)$; if the norm equals 1 for any such $p$, then the chain of inequalities in Eq. (11) cannot establish $\lim_{k\to\infty} \|D(k)\|_2 = 0$ by contraction alone. Alternatively, simulate two different input states through the same reservoir sequence and plot the trace distance per timestep: a step where the distance does not strictly decrease would contradict the contractivity claim.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that the iterative evolution of the homogenizer is governed by a single-step map $M(k) = p M_1^{(k)} + (1-p) M_S^{(k)}$, a convex combination of the identity channel and the SWAP channel, and that this map is contractive: for any two states, the $L^2$ distance between their images after one collision is bounded by the distance before the collision times a factor associated with $M(1)$, so repeated application sends the input state to the unique fixed point $\xi(0)$. Because contractivity is sufficient for asymptotic stability, the homogenizer satisfies the echo-state-type conditions required for temporal information processing. The paper emphasizes that the fixed point is an engineerable steady state rather than the maximally mixed state, which prevents the Volterra kernels from vanishing and gives the reservoir persistent memory.

Load-bearing premise

The entire proof of contractivity rests on the assumption that the single-step map $M = p\mathbb{1} + (1-p)\mathrm{SWAP}$ shrinks the $L^2$ distance between any two states by a factor strictly smaller than one.

Editorial extensions

If this is right

  • The homogenizer can process time-series data without precise Hamiltonian control: tuning the interaction time is enough to switch between fast convergence and reusable reservoir behavior.
  • Because the steady state is preparable and not maximally mixed, the reservoir retains non-vanishing memory, so temporal correlations encoded earlier continue to influence later outputs.
  • The same protocol works for arbitrarily long input sequences, since the argument is per-collision and the reservoir can be reused without reinitialization, at least in the weak-coupling regime.
  • The dissipative collision-model picture gives a concrete physical route to quantum reservoir computers on NMR and photonic hardware, with spin-based implementations possible.
  • A trade-off is identified: stronger coupling speeds convergence but shortens memory, while weaker coupling preserves reusability at the cost of slower homogenization.

Reading between the lines

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

  • A testable consequence of the convex decomposition $M = p\mathbb{1} + (1-p)\mathrm{SWAP}$ is that the convergence rate should be controlled by $(1-p)$ times the spectral gap of the SWAP channel; measuring the trace-distance decay for different $p$ would separate the role of the identity component from that of the swap component.
  • The contractivity argument, if it holds per collision, should generalize to higher-dimensional reservoir qudits and to multiple input qubits, since only the convexity of the map and the existence of a fixed point are used.
  • The same convergence-to-steady-state property suggests a secondary use: the homogenizer as a deterministic state-preparation and purification routine, independent of its role in machine learning.
  • The authors list NARMA benchmarks as future work; a direct way to test the reservoir-computing claim is to simulate the homogenizer on NARMA and compare its normalized mean-square error against classical echo-state networks with the same reservoir size.
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

4 major / 4 minor

Summary. The manuscript proposes the disordered quantum homogenizer, a collision-model-based system in which an input qubit interacts sequentially with a reservoir of identically prepared qubits via the partial SWAP operation, as a platform for quantum reservoir computing. The authors claim to prove that the homogenizer dynamics satisfies the stability and contractivity conditions (the quantum analog of the echo state property) by showing that the iterative CPTP map M = p M_1 + (1-p) M_S is contractive in the L2 norm, and they discuss physical implementations in NMR and photonic systems. The paper's central conclusion is that the homogenizer is a viable reservoir computer for temporal information processing.

Significance. The idea of using a simple, experimentally demonstrated homogenizer as a reservoir is attractive and would be a useful addition to the quantum reservoir computing literature if the mathematical claim were correct. The paper explicitly addresses the stability/contractivity criteria and cites relevant prior work. However, the central proof is not valid as written: the key inequality in Eq. (11) does not establish strict contraction, and the joint map M is in fact not strictly contractive in the L2 norm on the symmetric traceless subspace. Consequently, the paper's only new technical contribution, the proof of Definition 2, is not supported. The manuscript contains no learning benchmarks, so the practical claim rests entirely on the flawed proof.

major comments (4)
  1. [IV, Eq. (11)] The proof of contractivity is invalid. The step ∥M(k)(ξ(0)-ρ(0))∥2 ≤ ∥M(1)∥2·∥D(k-1)∥2 is a norm inequality, but the following equality '= M(1)∥D(k-1)∥2' replaces the operator norm by the symbol M(1) without justification, and no bound below 1 is established. For M = pM1 + (1-p)MS, the L2 (Hilbert-Schmidt) norm of the channel on traceless Hermitian operators is 1, not less than 1, because the identity component leaves symmetric traceless operators such as X = |00⟩⟨00| − |11⟩⟨11| invariant. Thus the single-step map is only non-expansive on the joint space, and the claimed strict contractivity in Definition 2 does not follow. The convergence of the reduced input state might be proven via the single-qubit channel Φ(ρ) = pρ + (1-p)ξ, which is strictly contractive with factor p, but that is not the distance considered in Eq. (11).
  2. [IV, after Eq. (11)] The sentence 'Clearly, by running the homogenizer long enough... we get lim_{k→∞}∥D(k)∥2 → 0' is unsupported by the preceding inequality. Non-expansiveness alone does not imply convergence to zero; a strict contraction factor or an independent convergence argument is required. The reference to [31] for fixed-point convergence (and to [53] for the mixing property) does not fill this gap, because the question is precisely whether the joint map is strictly contractive.
  3. [IV, paragraph after Eq. (8)] The proof of the central convergence statement relies on the authors' own prior work: 'as shown in [31]' and 'as we showed in [31]'. For a proof paper, this self-citation leaves a gap and raises a circularity concern, even though the original convergence result is externally grounded in Ziman et al. [29]. The argument should either be reproduced here or attributed explicitly to the original source.
  4. [Definitions 1 and 2] The definitions of stability and contractivity are not stated with sufficient precision for a proof. In Definition 1, the condition is written as lim_{N→∞}∥∏_{k=1}^N M(k)(ρ(0))∥2 = ξ(0), which is not a norm of a difference; it should read lim_{N→∞}∥∏ M(k)(ρ(0)) − ξ(0)∥2 = 0. In Definition 2, the distance function D(k) is introduced with the condition 'sup_{k∈Z} D(-k)(ξ(k) − ρ(k)) → 0', which is ill-typed (D(-k) is a sequence element, not a function). These imprecisions make the theorem statement difficult to verify.
minor comments (4)
  1. [Definition 1] The limit expression should be a norm of a difference, not ∥...∥2 = ξ.
  2. [Fig. 2] The caption lists J = 4 and J = 6, while the text and axis describe J = π/4 and J = π/6; please make the values consistent.
  3. [References] Reference [29] should cite the published version (Ziman et al., Phys. Rev. Lett. 88, 100405 (2002)) in addition to the arXiv identifier.
  4. [General] There are several typographical errors, e.g., 'Therefore,by' in the introduction and inconsistent spacing around citations.

Circularity Check

1 steps flagged · score 6.0 of 10

The contractivity proof in Eq. (11) assumes the contraction factor it is meant to establish, rendering the central derivation circular.

  1. self definitional [Section IV, Definition 2 and Eq. (11)]
    "To prove the satisfiability, let ∥D(k)∥2 ≡ ∥ξ(k) − ρ(k)ˆs∥2 ... ∥D(k)∥2 = ∥M(k)(ξ(0) − ρ(0)ˆs)∥2 ... ≤ ∥M(1)∥2 · ∥D(k−1)∥2 = M(1)∥D(k−1)∥2 (11) Clearly, by running the homogenizer long enough ... we get lim k→∞ ∥D(k)∥2 → 0"

    Definition 2 defines contractivity as the existence of a factor 0≤k<1 such that D(Mρ,Mξ)≤kD(ρ,ξ) and the distance sequence vanishes. In Eq. (11) the proof replaces the operator norm ∥M(1)∥2 by the scalar M(1) and then asserts the limit is zero. That replacement silently imports the contraction factor k=M(1)<1, which is exactly the contractivity statement to be proven. For the joint map M=p1+(1−p)SWAP, the identity component acts as the identity on the symmetric traceless subspace, so the joint map is only non-expansive, not strictly contractive; strict contraction holds only for the reduced single-qubit channel Φ(ρ)=pρ+(1−p)ξ, which the proof does not analyze. The derivation thus assumes its own conclusion.

full rationale

The only candidate circularity is the proof of contractivity in Section IV. Definition 2 makes contractivity equivalent to the existence of a contraction factor k<1 driving the distance sequence to zero. The proof's key identity in Eq. (11) converts the bound ∥M(1)∥2·∥D(k−1)∥2 into M(1)∥D(k−1)∥2 and then concludes convergence, thereby postulating the scalar contraction factor without deriving it from the channel. For the joint map M=p1+(1−p)SWAP the identity component leaves the symmetric traceless subspace invariant, so the operator norm is 1 and the strict factor is not available; a correct proof would need to analyze the reduced single-qubit map. This makes the central proof step circular: it assumes the map is contractive rather than demonstrating it. The self-citations to [31] are not treated as standalone circularity because the homogenizer's convergence is originally due to Ziman et al. [29] and [31] is a published, externally checkable result, so the underlying physical property has independent grounding. There is no fitted-parameter-named-prediction or renaming-of-known-result pattern. Overall score 6 reflects one load-bearing proof step that reduces to its own conclusion, while the physical phenomenon itself remains independently valid.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central proof rests almost entirely on assumptions rather than a self-contained derivation. The only new mathematical step is an unproven contraction assumption; the rest is imported from prior homogenizer and QRC results. No new entities are postulated.

assumptions (4)
  • ad hoc to paper The single-step CPTP map M is a strict contraction in the L2 norm with factor strictly less than 1.
    Eq. (11) concludes ||D(k)||2 -> 0 from the inequality ||D(k)||2 <= ||M(1)||2 ||D(k-1)||2, which requires ||M(1)||2 < 1. This is never proven, and for the stated channel M = pI + (1-p)SWAP the norm on symmetric traceless operators equals 1, so strict contractivity is not automatic.
  • domain assumption Reservoir qubits are independent and identically prepared in the canonical state xi at every relevant step.
    Eq. (1) and Definition 1 assume a product reservoir xi^{otimes N} and identical reservoir qubits, while the continuing temporal memory and reusability claims in Section IV require the reservoir to persist and evolve, which is in tension with the fresh-ancilla picture.
  • domain assumption The dynamics are input-independent and time-invariant, with bounded input sequences.
    Definition 1 requires input-independent CPTP maps with bounded real input sequences to obtain the echo-state property; the paper does not show how the time-varying input enters while keeping the map M input-independent.
  • domain assumption The homogenizer converges to a unique fixed point xi(0) for all initial input states.
    This convergence is imported from Refs. [29] and [31] and used in Eq. (8) and the contractivity argument; it is not re-derived in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dissipation-induced Quantum Homogenization for Temporal Information Processing." pith.science (2026). https://pith.science/paper/PGGU4ZOS

@misc{pith2026241209979,
  author       = {Pith},
  title        = {Pith review of: Dissipation-induced Quantum Homogenization for Temporal Information Processing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PGGU4ZOS}},
  note         = {Machine review of arXiv:2412.09979}
}
read the original abstract

Quantum reservoirs have great potential as they utilize the complex real-time dissipative dynamics of quantum systems for information processing and target time-series generation without precise control or fine-tuning of the Hamiltonian parameters. Nonetheless, their realization is challenging as quantum hardware with appropriate dynamics, robustness to noise, and ability to produce target steady states is required. To that end, we propose the disordered quantum homogenizer as an alternative platform, and prove it satisfies the necessary and sufficient conditions - stability and contractivity - of the reservoir dynamics, necessary for solving machine learning tasks with time-series input data streams. The results indicate that the quantum homogenization protocol, physically implementable as either nuclear magnetic resonance ensemble or a photonic system, can potentially function as a reservoir computer.

Figures

Figures reproduced from arXiv: 2412.09979 by the authors.

Figure 1
Figure 1. FIG. 1. Quantum circuit representation of the unitary trans [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Numerical simulation illustrating the convergence [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

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. Transient Dynamics and Homogenization in Incoherent Collision Models

    quant-ph 2025-01 conditional novelty 6.0 of 10

    In a collision model, replacing the coherent partial-swap with an incoherent controlled-swap preserves homogenization but suppresses transient memory effects and synchronization.

Reference graph

Works this paper leans on

64 extracted references · 53 canonical work pages · cited by 1 Pith paper

  1. [31]

    Ziman, P

    M. Ziman, P. Stelmachovic, V. Buzek, M. Hillery, V. Scarani, and N. Gisin, Quantum homogenization, arXiv:quant-ph/0110164

    arXiv

  2. [53]

    Reed and B

    M. Reed and B. Simon, Methods of modern mathemati- cal physics, Academic Press (1972)

    work page 1972

  3. [29]

    Negoro, K

    M. Negoro, K. Mitarai, K. Fujii, K. Nakajima, and 7 M. Kitagawa, Machine learning with controllable quan- tum dynamics of a nuclear spin ensemble in a solid, arXiv:1806.10910

    arXiv

  4. [1]

    Broughton, M

    HY Huang, M. Broughton, M. Mohseni, R. Babbush, S. Boixo, H. Neven, and J. McClean, Power of data in quan- tum machine learning, Nat. Commun. 12, 2631 (2021)

    work page 2021

  5. [2]

    Therefore, further supporting the idea that convergence is an extension of the mixing property

    which is sufficient to prove that the homogenizer has 8 Here, ξ(0) ∈ S(HR) can be viewed as the natural analog of the unique density operator, found in the mixing property [53], where as we showed in [31], the convergence is done via a noisy quantum channel, given by an input-independent CPTP map. Therefore, further supporting the idea that convergence is...

    work page

  6. [3]

    Hence, the distance between the input and reservoir density matri- ces (10) remains invariant

    If the coupling between the input and the kth reser- voir qubit is vanishing, J(t) ∼ 0, then ˜U ≡1, and no information will be encoded in that qubit. Hence, the distance between the input and reservoir density matri- ces (10) remains invariant. From a practical point of view, one could try to fix J(t) to a small non-zero value and simply increase the size...

    work page

  7. [4]

    Biamonte, P

    J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, and S. Lloyd, Quantum machine learning, Nature 549, 195–202 (2017)

    work page 2017

  8. [5]

    Barends, J

    R. Barends, J. Kelly, A. Megrant, A. Veitia, D. Sank, E. Jeffrey, T. White, J. Mutus, A. Fowler, B. Campbell, Y. Chen, Z. Chen, B. Chiaro, A. Dunsworth, C. Neill, P. O’Malley, P. Roushan, A. Vainsencher, J. Wenner, A. Korotkov, A. Cleland, and J. Martinis, Superconducting quantum circuits at the surface code threshold for fault tolerance, Nature 508, 500–...

    work page 2014

Show all 64 references
  1. [6]

    Horvat, X

    S. Horvat, X. Gao, and B. Daki´ c, Universal quantum computation via quantum controlled classical operations, J. Phys. A: Math. Th. 55, 075301 (2022)

    2022

  2. [7]

    Shor, Algorithms for quantum computation: discrete logarithms and factoring, Proceedings 35th Annual Sym- posium on Foundations of Computer Science (1994)

    P. Shor, Algorithms for quantum computation: discrete logarithms and factoring, Proceedings 35th Annual Sym- posium on Foundations of Computer Science (1994)

    1994

  3. [8]

    Anschuetz, J

    E. Anschuetz, J. Olson, A. Aspuru-Guzik, and Y. Cao, Variational quantum factoring, arXiv:1808.08927

    arXiv

  4. [9]

    Rønnow, Z

    T. Rønnow, Z. Wang, J. Job, S. Boixo, S. Isakov, D. Wecker, J. Martinis, D. Lidar, and M. Troyer, Defining and detecting quantum speedup, Science 345, 420 (2014)

    2014

  5. [10]

    Boixo, T

    S. Boixo, T. Rønnow, S. Isakov, Z. Wang, D. Wecker, D. Lidar, J. Martinis, and M. Troyer, Evidence for quantum annealing with more than one hundred qubits, Nat. Phys. 10, 218–224 (2014)

    2014

  6. [11]

    Farhi, J

    E. Farhi, J. Goldstone, S. Gutmann, J. Lapan, A. Lund- gren, and D. Preda, A quantum adiabatic evolution al- gorithm applied to random instances of an NP-complete problem, Science 292, 5516 (2001)

    2001

  7. [12]

    Fujii and S

    K. Fujii and S. Tamate, Computational quantum- classical boundary of noisy commuting quantum circuits, Sci. Rep. 6, 25598 (2016)

    2016

  8. [13]

    Stobinska, A

    M. Stobinska, A. Buraczewski, M. Moore, W. Clements, J. Renema, S. Nam, T. Gerrits, A. Lita, W. Koltham- mer, A. Eckstein, and I. Walmsley, Quantum interference enables constant-time quantum information processing, Science Adv. 5, 7 (2019)

    2019

  9. [14]

    Mujal, R

    P. Mujal, R. Mart ´ ınez-Pe˜ na, J. Nokkala, J. Garc ´ ıa-Beni, G. Giorgi, M. Soriano, and R. Zambrini, Opportunities in quantum reservoir computing and extreme learning machines, Adv. Quant. Tech., 2100027 (2021)

    2021

  10. [15]

    Fujii and K

    K. Fujii and K. Nakajima, Harnessing disordered- ensemble quantum dynamics for machine learning, Phys. Rev. Applied 8, 024030 (2017)

    2017

  11. [16]

    Chen and H

    J. Chen and H. Nurdin, Learning nonlinear input–output maps with dissipative quantum systems, Quant. Info. Proc. 18, 198 (2019)

    2019

  12. [17]

    Gallicchio and A

    C. Gallicchio and A. Micheli, Echo state property of deep reservoir computing networks, Cogn. Comp. 9 (2017)

    2017

  13. [18]

    Grigoryeva and J

    L. Grigoryeva and J. Ortega, Differentiable reservoir computing, JMLR 20, 1-62 (2019)

    2019

  14. [19]

    B. Li, R. Fong, and P. Tino, Simple cycle reservoirs are universal, JMLR 25, 1-28 (2024)

    2024

  15. [20]

    R. Fong, B. Li, and P. Tino, Universality of real minimal complexity reservoir, arXiv:2408.08071

    arXiv

  16. [21]

    Pena and J

    R. Pena and J. Ortega, Quantum reservoir computing in finite dimensions, Phys. Rev. E 107, 035306 (2023)

    2023

  17. [22]

    Fujii and K

    K. Fujii and K. Nakajima, Quantum reservoir computing: a reservoir approach toward quantum machine learning on near-term quantum devices, arXiv:2011.04890

    2011 arXiv

  18. [23]

    Kutvonen, K

    A. Kutvonen, K. Fujii, and T. Sagawa, Optimizing a quantum reservoir computer for time series prediction, Sci. Rep. 10, 14687 (2020)

    2020

  19. [24]

    Graves, A

    A. Graves, A. Mohamed, and G. Hinton, Speech recognition with deep recurrent neural networks, arXiv:1303.5778

    arXiv

  20. [25]

    Pascanu, C

    R. Pascanu, C. Gulcehre, K. Cho, and Y. Ben- gio, How to construct deep recurrent neural networks, arXiv:1312.6026

    arXiv

  21. [26]

    Jiang, Applications of deep learning in stock market prediction: recent progress, arXiv:2003.01859

    W. Jiang, Applications of deep learning in stock market prediction: recent progress, arXiv:2003.01859

    2003 arXiv

  22. [27]

    Verstraete, M

    F. Verstraete, M. Wolf, and J. Cirac, Quantum compu- tation and quantum-state engineering driven by dissipa- tion, Nat. Phys. 5, 633-636 (2009)

    2009

  23. [28]

    Alvarez-Rodriguez, L

    U. Alvarez-Rodriguez, L. Lamata, P. Escandell-Montero, J. Mart ´ ın-Guerrero, and E. Solano, Supervised quan- tum learning without measurements, Sci. Rep. 7, 13645 (2017)

    2017

  24. [30]

    Violaris, G

    M. Violaris, G. Bhole, J. Jones, V. Vedral, and C. Mar- letto, Transforming pure and mixed states using an NMR quantum homogenizer, Phys. Rev. A 103, 022414 (2021)

    2021

  25. [32]

    Marletto, V

    C. Marletto, V. Vedral, L. Knoll, F. Piacentini, E. Bernardi, E. Rebufello, A. Avella, and M. Gramegna, Emergence of constructor-based irreversibility in quan- tum systems: theory and experiment, Phys. Rev. Lett. 128, 080401 (2022)

    2022

  26. [33]

    Yosifov, A

    A. Yosifov, A. Iyer, D. Ebler, and V. Vedral, Quantum homogenization as a quantum steady-state protocol on noisy intermediate-scale quantum hardware, Phys. Rev. A 109, 032624 (2024)

    2024

  27. [34]

    Volya and P

    D. Volya and P. Mishra, State preparation on quantum computers via quantum steering, IEEE Trans. Quant. Eng. 5, 3100714 (2024)

    2024

  28. [35]

    Cattaneo, G

    M. Cattaneo, G. Chiara, S. Maniscalco, R. Zambrini, and G. Giorgi, Collision models can efficiently simulate any multipartite Markovian quantum dynamics, Phys. Rev. Lett. 126, 130403 (2021)

    2021

  29. [36]

    Ciccarello, G

    F. Ciccarello, G. Palma, and V. Giovannetti, Collision- model-based approach to non-Markovian quantum dy- namics, Phys. Rev. A 87, 040103(R) (2013)

    2013

  30. [37]

    Dudas, B

    J. Dudas, B. Carles, E. Plouet, F. Mizrahi, J. Grol- lier, and D. Markovi´ c, Quantum reservoir computing im- plementation on coherently coupled quantum oscillators, npj Quant. Info. 9, 64 (2023)

    2023

  31. [38]

    Govia, G

    L. Govia, G. Ribeill, G. Rowlands, and T. Ohki, Non- linear input transformations are ubiquitous in quantum reservoir computing, Neuromorph. Comp. Eng. 2, 014008 (2022)

    2022

  32. [39]

    T. Ali, A. Bhattacharyya, S. Haque, E. Kim, N. Moyni- han, and J. Murugan, Chaos and complexity in quantum mechanics, Phys. Rev. D 101, 026021 (2020)

    2020

  33. [40]

    Kane, A silicon-based nuclear spin quantum computer, Nature 393, 133–137 (1998)

    B. Kane, A silicon-based nuclear spin quantum computer, Nature 393, 133–137 (1998)

    1998

  34. [41]

    DiVincenzo, D

    D. DiVincenzo, D. Bacon, J. Kempe, G. Burkard, and K. Whaley, Universal quantum computation with the ex- change interaction, Nature 408, 339–342 (2000)

    2000

  35. [42]

    Dambre, D

    J. Dambre, D. Verstraeten, B. Schrauwen, and S. Massar, Information processing capacity of dynamical systems, Sci. Rep. 2, 514 (2012)

    2012

  36. [43]

    Kobayashi, Q

    S. Kobayashi, Q. Tran, and K. Nakajima, Extending echo state property for quantum reservoir computing, Phys. Rev. E 110, 024207 (2024)

    2024

  37. [44]

    Kobayashi, Q

    S. Kobayashi, Q. Tran, and K. Nakajima, Coherence in- flux is indispensable for quantum reservoir computing, arXiv:2409.12693

    arXiv

  38. [45]

    Sandberg, Nonlinear input-output maps and approx- imate representations, AT&T Technical Journal 64, 8 (1985)

    I. Sandberg, Nonlinear input-output maps and approx- imate representations, AT&T Technical Journal 64, 8 (1985)

    1985

  39. [46]

    Hanson and M

    J. Hanson and M. Raginsky, Universal approximation of input-output maps by temporal convolutional nets, arXiv:1906.09211

    1906 arXiv

  40. [47]

    Gallicchio and A

    C. Gallicchio and A. Micheli, Architectural and Marko- vian factors of echo state networks, Neural Networks 24, 5 (2011)

    2011

  41. [48]

    echo state

    H. Jaeger, The “echo state” approach to analysing and training recurrent neural networks, German National Re- search Center for Information Technology (2010)

    2010

  42. [49]

    B. Kia, A. Mendes, A. Parnami, R. George, K. Mob- ley, and W. Ditto, Nonlinear dynamics based machine learning: Utilizing dynamics-based flexibility of nonlin- ear circuits to implement different functions, PLOS ONE 15(3), 0228534 (2020)

    2020

  43. [50]

    Jaeger, M

    H. Jaeger, M. Lukoˇ seviˇ cius, D. Popovici, U. Siewert, Op- timization and applications of echo state networks with leaky- integrator neurons, Neural Networks 20, 3 (2007)

    2007

  44. [51]

    Rudin, Principles of Mathematical Analysis (1964)

    W. Rudin, Principles of Mathematical Analysis (1964)

    1964

  45. [52]

    Searcoid, Metric spaces, Springer Science and Busi- ness Media (2006)

    M. Searcoid, Metric spaces, Springer Science and Busi- ness Media (2006)

    2006

  46. [54]

    Raginsky, Strictly contractive quantum channels and physically realizable quantum computers, Phys

    M. Raginsky, Strictly contractive quantum channels and physically realizable quantum computers, Phys. Rev. A 65, 032306 (2002)

    2002

  47. [55]

    Burgarth, G

    D. Burgarth, G. Chiribella, V. Giovannetti, P. Perinotti, and K. Yuasa, Ergodic and mixing quantum channels in finite dimensions, New J. Phys. 15, 073045 (2013)

    2013

  48. [56]

    Beever, M

    A. Beever, M. Violaris, C. Marletto, and V. Vedral, Com- paring coherent and incoherent models for quantum ho- mogenization, Phys. Rev. A 110, 012464 (2024)

    2024

  49. [57]

    Buehner and P

    M. Buehner and P. Young, A tighter bound for the echo state property, IEEE Trans. on Neural Networks 17, 3 (2006)

    2006

  50. [58]

    J. Chen, H. Nurdin, and N. Yamamoto, Temporal infor- mation processing on noisy quantum computers, Phys. Rev. A 14, 024065 (2020)

    2020

  51. [59]

    Yasuda, Y

    T. Yasuda, Y. Suzuki, T. Kubota, K. Nakajima, Q. Gao, W. Zhang, S. Shimono, H. Nurdin, and N. Yamamoto, Quantum reservoir computing with repeated measure- ments on superconducting devices, arXiv:2310.06706

    arXiv

  52. [60]

    F. Hu, S. Khan, N. Bronn, G. Angelatos, G. Rowlands, G. Ribeill, and H. T¨ ureci, Overcoming the coherence time barrier in quantum machine learning on temporal data, Nat. Commun. 15, 7491 (2024)

    2024

  53. [61]

    Tran and K

    Q. Tran and K. Nakajima, Higher-order quantum reser- voir computing, arXiv:2006.08999

    2006 arXiv

  54. [62]

    Jozsa, Fidelity for mixed quantum states, J

    R. Jozsa, Fidelity for mixed quantum states, J. Mod. Opt. 41, 2315 (1994)

    1994

  55. [63]

    Gallicchio and A

    C. Gallicchio and A. Micheli, Deep echo state network (DeepESN): a brief survey, arXiv:1712.04323

    arXiv

  56. [64]

    Levy, Universal quantum computation with spin-1/2 pairs and Heisenberg exchange, Phys

    J. Levy, Universal quantum computation with spin-1/2 pairs and Heisenberg exchange, Phys. Rev. Lett. 89, 147902 (2002)

    2002

Pith tools

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