Pith. sign in

REVIEW 3 major objections 5 minor 68 references

Memory-minimal quantum generation of stochastic processes: spectral invariants of quantum hidden Markov models

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 ↗ Review JSON

Pith's one-line read The spectrum of a process's transfer operator is the same in every generating quantum hidden Markov model, and it fixes a lower bound on the memory needed to generate the process.

desk verdict The spectral invariant at the core of this paper is false as stated; the paper is worth a referee only if the authors restrict to contributing eigenvalues. read the letter →

arxiv 2412.12812 v1 pith:AOG74JWM submitted 2024-12-17 quant-ph

Magdalini Zonnios , Alec Boyd , Felix C. Binder This is my paper · ORCID
classification quant-ph MSC 81P4560J10 PACS 03.67.-a02.50.Ga
keywords quantumhiddenMarkovmodelsstochasticprocessesgenerativecomplexitytransferoperatorspectralinvariantscoherencestrictlyincoherentoperationsmemory-minimal
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 aims to solve part of the problem of telling how much hidden memory a stochastic process really needs: infinitely many quantum hidden Markov models can produce the same observed statistics, and before this paper there was no known feature of the process itself that all generating models must share. It identifies such a feature: the set of distinct nonzero eigenvalues of the transfer operator built from any generating model is the same for every model of that process. From this invariant it derives a lower bound on the quantum generative topological complexity, and shows that forbidding quantum coherence raises the bound quadratically, so any model that beats the classical bound must use coherence. This matters because it turns a seemingly uncomputable minimization over infinitely many models into a bound that can be read off from one model, and it gives a precise sense in which quantum memory advantages are real.

What carries the argument

The transfer operator $E_{AB}=\sum_x \mathcal{E}^x_A\otimes\mathcal{E}^x_B$ and its vectorised form carry the argument. Contracting the outputs of two copies of the instrument and taking the $L$-th power makes the probability of every length-$L$ word equal to $\langle\!\langle 1|E_{AB}^L|\rho_{AB}\rangle\!\rangle$, so two models generate the same process exactly when these traces agree for all $L$. The distinct nonzero eigenvalues of $E_{AB}$ then become a fingerprint of the process rather than of any particular model: equality for all $L$ forces the spectra of any two generating models to coincide. The dimension counting that produces the bounds is simply that an $n$-dimensional memory makes the vectorised transfer operator an $n^4\times n^4$ matrix, while a classical strictly incoherent model makes that matrix sparse with rank at most $n^2$.

What would settle it

Compute the distinct nonzero spectra of the vectorised transfer operators of two QHMMs that generate the same stochastic process, counting only eigenvalues whose $\alpha_\lambda$ is nonzero; if the two sets differ, Theorem 1 fails. A more direct test is to append a decoupled two-level classical sector to a known minimal QHMM: the visible process is unchanged, and checking whether the resulting $|\Lambda|$ changes reveals whether the invariant needs the nonzero-coefficient caveat.

Watch

Extended reading notes

Core claim

The paper's central claim is that every quantum hidden Markov model that generates a given stochastic process shares one spectral invariant: the set of distinct nonzero eigenvalues $\Lambda$ of the vectorised transfer operator $E_{AB}=\sum_x \mathcal{E}^x_A\otimes \mathcal{E}^x_B$, together with the coefficients $\alpha_\lambda=\langle\!\langle 1|\Pi_\lambda|\rho\rangle\!\rangle$ for shared eigenvalues. Because the transfer operator is an $n^4\times n^4$ matrix when the memory has dimension $n$, a process whose invariant spectrum has $|\Lambda|$ distinct nonzero eigenvalues cannot be generated in fewer than $\log\lceil |\Lambda|^{1/4}\rceil$ memory units, giving Theorem 2. When the generating operations are restricted to strictly incoherent operations, the quantum description of a classical hidden Markov model, the same invariant requires $\log\lceil |\Lambda|^{1/2}\rceil$ memory, a quadratically larger bound. Since the quantum bound is weaker, the gap between the two bounds is a resource-theoretic signature of coherence, and Theorem 4 exhibits a three-state classical process whose emissions fit in a two-dimensional quantum memory, proving $c_Q<c_C$ for that process.

Load-bearing premise

The load-bearing premise is that every distinct nonzero eigenvalue in $\Lambda$ actually contributes to the process with a nonzero coefficient; if a model has a decoupled memory sector whose eigenvalues have coefficient zero, the invariant as stated counts eigenvalues that do not affect the emitted statistics, which can push the lower bound above the true minimal memory.

Editorial extensions

If this is right

  • The lower bound $c_Q(\vec{X})\ge \log\lceil |\Lambda|^{1/4}\rceil$ can be evaluated from any valid QHMM, so even a non-minimal model certifies a floor on the memory needed by any generator of the process.
  • For strictly incoherent (classical) generation the bound becomes $c_C(\vec{X})\ge \log\lceil |\Lambda|^{1/2}\rceil$, so any quantum model that beats the classical bound must exploit coherence relative to the measurement basis.
  • There are concrete processes, such as the three-state example in the paper, with $c_C(\vec{X})=\log 3$ and $c_Q(\vec{X})=\log 2$, demonstrating a strict quantum advantage in generative memory.
  • Any operation applied to a QHMM that leaves the generated process unchanged must preserve the distinct nonzero spectrum of its transfer operator, which sharply constrains the set of alternative presentations.

Reading between the lines

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

  • Inference: the invariant is best read as the set of distinct nonzero eigenvalues with nonzero coefficients $\alpha_\lambda=\langle\!\langle 1|\Pi_\lambda|\rho\rangle\!\rangle$. A nonminimal model with an unreachable or decoupled memory sector contributes eigenvalues with $\alpha_\lambda=0$, and the paper's Appendix A itself concedes the equality of spectra only 'unless $\alpha_\lambda=0$'; stated
  • Inference: the fourth-root counting suggests a direct algebraic characterisation of minimal quantum dimension as the smallest $d$ for which some $d^4$-dimensional vectorised transfer operator realises the invariant spectrum; an algorithm that factors a candidate transfer operator into a tensor-product instrument structure would turn the bound into a constructive minimality test.
  • Inference: because strictly incoherent transfer operators have rank at most $m^2$, one can search for quantum memory advantages by comparing the numerical rank of a candidate quantum transfer operator with the rank required by any classical edge-emitting HMM for the same word probabilities, a finite linear-algebra probe.
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 proposes spectral invariants for quantum hidden Markov models (QHMMs). It defines the vectorized transfer operator E_AB = Σ_x E_x ⊗ E_x, claims in Theorem 1 that the set Λ of distinct nonzero eigenvalues of this operator is invariant across all QHMMs generating the same stochastic process, and uses this claim to lower-bound the generative topological complexity: c_Q ≥ log⌈|Λ|^{1/4}⌉ in Theorem 2 and c_C ≥ log⌈|Λ|^{1/2}⌉ for classically generated (strictly incoherent) processes in Theorem 3. Theorem 4 then constructs a three-state classical HMM that admits a two-level quantum realization, claimed to show c_Q < c_C. The appendices contain the proofs and the explicit construction of the quantum model from a classical HMM.

Significance. If correct, the spectral-invariant approach would be a genuinely useful tool: minimality of generative models is generally hard to certify, and a computable invariant from any non-minimal model would give rigorous memory lower bounds. The tensor-network formulation of QHMMs, the resource-theoretic framing in terms of coherence, and the explicit constructive example of a quantum memory advantage are appealing features. The paper is clearly written and the main ideas are easy to identify. However, the central invariant as stated is false, and Theorem 2 and Theorem 3 inherit this flaw. The underlying idea appears salvageable by restricting to eigenvalues with nonzero coefficient α_λ, but the manuscript as written does not state, prove, or use that restriction. The significance of the contribution is therefore conditional on a substantial correction of the main theorem and its consequences.

major comments (3)
  1. [Theorem 1 and Appendix A (Eq. (7), Eqs. (A9)–(A19))] Theorem 1 is false as stated. The Vandermonde argument in Appendix A does not show that the symmetric difference Λ_R Δ Λ_Q is empty; it shows only that any eigenvalue appearing in exactly one model must have zero coefficient α_λ. The appendix itself concedes this with the phrase "unless α_λ_R = 0 or α_λ_Q = 0". This is not a harmless edge case. For the i.i.d. fair-coin process, a one-state model with T^(0)=T^(1)=1/2 has transfer-operator eigenvalue set {1/2}, while the two-state model with T^(0)=diag(1/2,0.3), T^(1)=diag(1/2,0.7), and initial distribution π=(1,0) generates the same fair-coin process but has transfer operator E=Σ_x T^(x)⊗T^(x) with distinct nonzero eigenvalues {1/2,0.58}; the extra eigenvalue has α=0 because the initial state has no support on the second decoupled sector. Thus Eq. (7), the claimed process invariance of Λ, is false. The theorem also silently assumes diagonalizability of the transfer operator; footnote [60] asserting that non-diagonalizable matrices are measure zero is not a proof that all valid QHMM transfer operators are diagonalizable.
  2. [Theorems 2 and 3 (Eqs. (11)–(12))] Because Theorem 1 is false, the bounds in Theorems 2 and 3 are not valid for an arbitrary generating model. In the fair-coin counterexample above, applying Theorem 2 to the two-state model would give c_Q ≥ log⌈2^{1/4}⌉ = log 2, while the fair-coin process has c_Q = 0. The source of the error is that Λ, as defined, includes eigenvalues that are in the spectrum of the transfer operator but do not contribute to any sequence probability because their coefficient α_λ = ⟨⟨1|Π_λ|ρ⟩⟩ vanishes. The proofs must be revised to use the "active spectrum" S = {λ ∈ Λ : α_λ ≠ 0}, with a proof that S is invariant across generating models. Without this correction, the claimed lower bounds in Eqs. (11) and (12) do not follow.
  3. [Theorem 4 and its proof (Eq. (14), Appendix E)] The proof of Theorem 4 is incomplete as written. The three-state classical model is asserted to be "classically irreducible", but no argument is given that no two-state classical HMM generates the same process. The text says "using Thm. 3 we can verify c_C = log 3", but Theorem 3 is only a lower bound, and no value of |Λ| for the example is actually computed. Similarly, the construction gives a two-dimensional quantum memory, so c_Q ≤ log 2, but to conclude c_Q = log 2 one must exclude one-dimensional memories; this is not shown. The authors should supply the missing computation of the invariant (with the corrected α_λ ≠ 0 definition) and a direct minimality argument for the classical model.
minor comments (5)
  1. [Introduction] There is a duplicated word: "any any valid presentation" near the end of the introduction.
  2. [Proof of Theorem 4 / Appendix E] There are several typos: "memeory reduction" and "memeory states" should be "memory reduction" and "memory states".
  3. [Eq. (14)] The definition "¯⋆ := 1−⋆1−⋆2" appears to be a typo; it should presumably be "¯⋆ := 1−⋆".
  4. [Appendix D, Eq. (D8)] The displayed inequality "c_C(−→X)^2 ≥ log[rank(E_RCmin)] ≥ log |Λ|" appears dimensionally wrong: since c_C = log m and rank ≤ m^2, the correct inequality is 2 c_C ≥ log rank, not c_C^2 ≥ log rank. The final bound in Eq. (12) is correct, but the proof as written needs this corrected.
  5. [Theorem 3 proof (main text, after Eq. (13))] The sentence "any minimal QHMM (without the SIO constraint) cannot exceed the classical HMM's memory since this would contradict Thm. (2)" is not a valid justification for c_C ≥ c_Q. The inequality c_C ≥ c_Q follows simply because every classical HMM is a QHMM, so the minimum in Eq. (2) can only be smaller.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the spectral-bound argument is a forward mathematical derivation from model definitions; the paper's known gap in Theorem 1 is an overstatement, not a circular reduction.

full rationale

The central derivation is self-contained: QHMMs are defined by Eqs. (1)-(2), the transfer operator is introduced in Eq. (4), and the claimed spectral invariant is argued directly from the equal-process condition via the Vandermonde argument in Appendix A. Theorems 2 and 3 then combine that invariant with dimensional counting on the vectorized transfer operator. No parameter is fitted to data and renamed a prediction, no empirical benchmark is used as input, and no uniqueness result is imported from the authors' prior work to exclude alternative models. The few self-citations (e.g., refs. [26], [55], [56], [65]) appear as background context or notational references, not as load-bearing premises. The paper's substantive weakness is mathematical rather than circular: Appendix A itself concedes, after Eq. (A19), that equality of spectra holds 'unless α_λR = 0 or α_λQ = 0 for any eigenvalue λR ∈ ΛR or λQ ∈ ΛQ'. The proof therefore does not establish that every nonzero eigenvalue of an arbitrary generating transfer operator belongs to the process invariant, because eigenvalues with zero coefficient can be added by decoupled memory sectors. This is a correctness gap in the stated bound, not a case where the conclusion is equivalent to the input by construction.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The derivation relies on the transfer-operator spectrum as a lower-bound certificate. The main free choices are the example parameters; the axioms include the unstated nonzero-coefficient condition and the restricted SIO rank structure, both load-bearing.

free parameters (2)
  • α = 0.5
    Chosen by hand in the example (Theorem 4) to exhibit the quantum memory advantage; the HMM is parameterized by α and β.
  • β
    Parameter of the example HMM; left unspecified, varied to trace the advantage region.
assumptions (5)
  • domain assumption The transfer operator E_RR is diagonalisable for the models considered.
    Invoked to write E = Σ λ Π_λ and to apply the Vandermonde argument; footnote 60 says non-diagonalisable matrices are measure zero.
  • domain assumption Two models generate the same process iff the moment sequences <<1|E^L|ρ>> are equal for all L.
    Used in Eq. (6) as the starting point of Theorem 1.
  • ad hoc to paper Every eigenvalue in Λ has a nonzero coefficient α_λ, or such eigenvalues can be ignored.
    Needed for the conclusion Λ_R=Λ_Q; the proof only shows equality for eigenvalues with α≠0, and the theorem statement omits this condition.
  • domain assumption SIO transfer operators have nonzero entries only for diagonal index pairs, giving rank ≤ m^2.
    Used in Theorem 3 (Eq. C5); true for the dephasing-type Kraus representation K_{i→j}=sqrt(p)|j><i|, but not for all SIOs such as permutation unitaries that preserve coherences.
  • domain assumption The two-dimensional QHMM constructed via non-orthogonal states in Theorem 4 is a valid CPTP instrument generating the same process.
    The main text asserts the reduction without explicitly giving the instrument; Appendix E provides a general recipe, but the specific parameters are not checked.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Memory-minimal quantum generation of stochastic processes: spectral invariants of quantum hidden Markov models." pith.science (2026). https://pith.science/paper/AOG74JWM

@misc{pith2026241212812,
  author       = {Pith},
  title        = {Pith review of: Memory-minimal quantum generation of stochastic processes: spectral invariants of quantum hidden Markov models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AOG74JWM}},
  note         = {Machine review of arXiv:2412.12812}
}
read the original abstract

Stochastic processes abound in nature and accurately modeling them is essential across the quantitative sciences. They can be described by hidden Markov models (HMMs) or by their quantum extensions (QHMMs). These models explain and give rise to process outputs in terms of an observed system interacting with an unobserved memory. Although there are infinitely many models that can generate a given process, they can vary greatly in their memory requirements. It is therefore of great fundamental and practical importance to identify memory-minimal models. This task is complicated due to both the number of generating models, and the lack of invariant features that determine elements of the set. In general, it is forbiddingly difficult to ascertain that a given model is minimal. Addressing this challenge, we here identify spectral invariants of a process that can be calculated from any model that generates it. This allows us to determine strict bounds on the quantum generative complexity of the process -- its minimal memory requirement. We then show that the bound is raised quadratically when we restrict to classical operations. This is an entirely quantum-coherent effect, as we express precisely, using the resource theory of coherence. Finally, we demonstrate that the classical bound can be violated by quantum models.

Figures

Figures reproduced from arXiv: 2412.12812 by the authors.

Figure 1
Figure 1. The transfer operator EP P ′ = P x EP ⊗ EP ′ (high￾lighted in the purple box) is obtained by contracting the out￾put of two tensor chains for processes P and P ′ which are obtained by a sequence of quantum instruments acting on memory. ing QHMMs as tensor chains, we show that the distinct, non-zero spectrum of their transfer operators serve as a process invariant, allowing us to derive a lower bound on cQ( −→X) whic… view at source ↗
Figure 2
Figure 2. The inset shows a classically irreducible three-state, [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. a) The transfer operator of a QHMM can be obtained as [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

68 extracted references · 58 canonical work pages

  1. [60]

    Non-diagonalisable matrices are measure zero in the space of randomly matrices, representing a dense set of stochastic processes

    work page

  2. [1]

    Rabiner, Proceedings of the IEEE 77, 257 (1989), con- ference Name: Proceedings of the IEEE

    L. Rabiner, Proceedings of the IEEE 77, 257 (1989), con- ference Name: Proceedings of the IEEE

    work page 1989

  3. [2]

    D. R. Upper, Theory and algorithms for hidden Markov models and generalized hidden Markov models, Ph.D. the- sis, University of California, Berkeley (1997)

    work page 1997

  4. [3]

    M. O. Rabin, Information and Control 6, 230 (1963)

    work page 1963

  5. [4]

    Gruska, D

    J. Gruska, D. Qiu, and S. Zheng, Int. J. Found. Comput. Sci. 26, 381 (2015)

    work page 2015

  6. [5]

    Y. Tian, T. Feng, M. Luo, S. Zheng, and X. Zhou, npj Quantum Inf 5, 1 (2019)

    work page 2019

  7. [6]

    Giovannetti, S

    V. Giovannetti, S. Lloyd, and L. Maccone, Phys. Rev. Lett. 100, 160501 (2008)

    work page 2008

  8. [7]

    Wiesner and J

    K. Wiesner and J. P. Crutchfield, Physica D: Nonlinear Phenomena Novel Computing Paradigms: Quo Vadis?, 237, 1173 (2008)

    work page 2008

Show all 68 references
  1. [8]

    O’Neill, T

    B. O’Neill, T. M. Barlow, D. ˇSafr´ anek, and A. Beige, AIP Conference Proceedings 1479, 667 (2012)

    2012

  2. [9]

    L. A. Clark, W. Huang, T. M. Barlow, and A. Beige, in ISCS 2014: Interdisciplinary Symposium on Com- plex Systems , edited by A. Sanayei, O. E. R¨ ossler, and I. Zelinka (Springer International Publishing, Cham,

    2014

  3. [10]

    Monr` as and A

    A. Monr` as and A. Winter, Journal of Mathematical Physics 57, 015219 (2016)

    2016

  4. [11]

    Cholewa, P

    M. Cholewa, P. Gawron, P. G lomb, and D. Kurzyk, Quantum Inf Process 16, 101 (2017)

    2017

  5. [12]

    Srinivasan, G

    S. Srinivasan, G. Gordon, and B. Boots, in Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics (PMLR, 2018) pp. 1979–1987

    2018

  6. [13]

    Aloisio, G

    I. Aloisio, G. White, C. Hill, and K. Modi, PRX Quan- tum 4, 020310 (2023)

    2023

  7. [14]

    M. Gu, K. Wiesner, E. Rieper, and V. Vedral, Nat Com- mun 3, 762 (2012)

    2012

  8. [15]

    J. R. Mahoney, C. Aghamohammadi, and J. P. Crutch- field, Scientific Reports 6, 20495 (2016)

    2016

  9. [16]

    P. M. Riechers, J. R. Mahoney, C. Aghamohammadi, and J. P. Crutchfield, Phys. Rev. A 93, 052317 (2016)

    2016

  10. [17]

    Aghamohammadi, J

    C. Aghamohammadi, J. R. Mahoney, and J. P. Crutch- field, Sci Rep 7, 6735 (2017)

    2017

  11. [18]

    A. P. Lund, M. J. Bremner, and T. C. Ralph, npj Quan- tum Inf 3, 15 (2017)

    2017

  12. [19]

    A. J. P. Garner, Q. Liu, J. Thompson, V. Vedral, and mile Gu, New Journal of Physics 19, 103009 (2017)

    2017

  13. [20]

    T. J. Elliott and M. Gu, npj Quantum Inf 4, 1 (2018)

    2018

  14. [21]

    Schuld and F

    M. Schuld and F. Petruccione, in Supervised Learn- ing with Quantum Computers , edited by M. Schuld and F. Petruccione (Springer International Publishing, Cham, 2018) pp. 247–272

    2018

  15. [22]

    S. P. Loomis and J. P. Crutchfield, J Stat Phys176, 1317 (2019)

    2019

  16. [23]

    Ghafari, N

    F. Ghafari, N. Tischler, J. Thompson, M. Gu, L. K. Shalm, V. B. Verma, S. W. Nam, R. B. Patel, H. M. Wiseman, and G. J. Pryde, Phys. Rev. X 9, 041013 (2019)

    2019

  17. [24]

    T. J. Elliott, A. J. P. Garner, and M. Gu, New J. Phys. 21, 013021 (2019)

    2019

  18. [25]

    M. Ho, M. Gu, and T. J. Elliott, Phys. Rev. A 101, 032327 (2020)

    2020

  19. [26]

    T. J. Elliott, C. Yang, F. C. Binder, A. J. Garner, J. Thompson, and M. Gu, Phys. Rev. Lett. 125, 260501 (2020)

    2020

  20. [27]

    Blank, D

    C. Blank, D. K. Park, and F. Petruccione, npj Quantum Inf 7, 1 (2021)

    2021

  21. [28]

    Korzekwa and M

    K. Korzekwa and M. Lostaglio, Phys. Rev. X 11, 021019 (2021)

    2021

  22. [29]

    T. J. Elliott, M. Gu, A. J. Garner, and J. Thompson, Phys. Rev. X 12, 011007 (2022)

    2022

  23. [30]

    K.-D. Wu, C. Yang, R.-D. He, M. Gu, G.-Y. Xiang, C.-F. Li, G.-C. Guo, and T. J. Elliott, Nat Commun 14, 2624 (2023)

    2023

  24. [31]

    Hangleiter and J

    D. Hangleiter and J. Eisert, Rev. Mod. Phys. 95, 035001 (2023)

    2023

  25. [32]

    Banchi, Machine Learning: Science and Technology 5 (2024), 10.1088/2632-2153/ad444a

    L. Banchi, Machine Learning: Science and Technology 5 (2024), 10.1088/2632-2153/ad444a

    2024 doi

  26. [33]

    T. J. Elliott and M. Gu, Phys. Rev. A 109, 022434 (2024)

    2024

  27. [34]

    C. J. Ellison, J. R. Mahoney, R. G. James, J. P. Crutch- field, and J. Reichardt, Chaos: An Interdisciplinary Journal of Nonlinear Science 21 (2011)

    2011

  28. [35]

    C. R. Shalizi and J. P. Crutchfield, Journal of Statistical Physics 104, 817 (2001)

    2001

  29. [36]

    Ay and J

    N. Ay and J. P. Crutchfield, J Stat Phys 120, 659 (2005)

    2005

  30. [37]

    Hidden Quantum Markov Models and non-adaptive read-out of many-body states,

    A. Monras, A. Beige, and K. Wiesner, “Hidden Quantum Markov Models and non-adaptive read-out of many-body states,” (2012), arXiv:1002.2337 [quant-ph]

    2012 arXiv

  31. [38]

    J. B. Ruebeck, R. G. James, J. R. Mahoney, and J. P. Crutchfield, Chaos: An Interdisciplinary Journal of Non- linear Science 28, 013109 (2018)

    2018

  32. [39]

    A. M. Jurgens and J. P. Crutchfield, Chaos: An Interdis- ciplinary Journal of Nonlinear Science 31, 083114 (2021)

    2021

  33. [40]

    Dimension reduction in quantum sampling of stochastic processes,

    C. Yang, M. Florido-Llin‘as, M. Gu, and T. J. Elliott, “Dimension reduction in quantum sampling of stochastic processes,” (2024), arXiv:2404.10338 [quant-ph]

    2024 arXiv

  34. [41]

    L¨ ohr, J Syst Sci Complex25, 30 (2012)

    W. L¨ ohr, J Syst Sci Complex25, 30 (2012)

    2012

  35. [42]

    Stokes and J

    J. Stokes and J. Terilla, Entropy 21, 1236 (2019)

    2019

  36. [43]

    Glasser, R

    I. Glasser, R. Sweke, N. Pancotti, J. Eisert, and J. I. 6 Cirac, in Neural Information Processing Systems (2019)

    2019

  37. [44]

    T. Li, D. Precup, and G. Rabusseau, Mach Learn 113, 2619 (2024)

    2024

  38. [45]

    Glasser, N

    I. Glasser, N. Pancotti, and J. I. Cirac, IEEE Access 8, 68169 (2020)

    2020

  39. [46]

    Khavari and G

    B. Khavari and G. Rabusseau, CoRR abs/2106.11827 (2021), 2106.11827

    2021 arXiv

  40. [47]

    M. L. Wall and G. D’Aguanno, Phys. Rev. A104, 042408 (2021)

    2021

  41. [48]

    Tensor networks and efficient descriptions of classical data,

    S. Lu, M. Kan´ asz-Nagy, I. Kukuljan, and J. I. Cirac, “Tensor networks and efficient descriptions of classical data,” (2021), version Number: 1

    2021

  42. [49]

    Adhikary, S

    S. Adhikary, S. Srinivasan, J. Miller, G. Rabusseau, and B. Boots, PMLR 130, 2080 (2021)

    2021

  43. [50]

    The resource theory of tensor networks,

    M. Christandl, V. Lysikov, V. Steffan, A. H. Werner, and F. Witteveen, “The resource theory of tensor networks,” (2023), arXiv:2307.07394 [quant-ph]

    2023 arXiv

  44. [51]

    Sequence Processing with Quantum Tensor Networks,

    C. Harvey, R. Yeung, and K. Meichanetzidis, “Sequence Processing with Quantum Tensor Networks,” (2023), arXiv:2308.07865 [quant-ph]

    2023 arXiv

  45. [52]

    Rieser, F

    H.-M. Rieser, F. K¨ oster, and A. P. Raulf, Proc. R. Soc. A. 479, 20230218 (2023)

    2023

  46. [53]

    H. Liao, I. Convy, Z. Yang, and K. B. Whaley, Quantum Mach. Intell. 5, 7 (2023)

    2023

  47. [54]

    Efficient quantum mixed-state tomography with un- supervised tensor network machine learning,

    W. jun Li, K. Xu, H. Fan, S. ju Ran, and G. Su, “Efficient quantum mixed-state tomography with un- supervised tensor network machine learning,” (2023), arXiv:2308.06900 [quant-ph]

    2023 arXiv

  48. [55]

    F. C. Binder, J. Thompson, and M. Gu, Phys. Rev. Lett. 120, 240502 (2018)

    2018

  49. [56]

    Q. Liu, T. J. Elliott, F. C. Binder, C. Di Franco, and M. Gu, Phys. Rev. A 99, 062110 (2019)

    2019

  50. [57]

    T. J. Elliott, Phys. Rev. A 103, 052615 (2021)

    2021

  51. [58]

    L. B. Vieira and C. Budroni, Quantum 6, 623 (2022)

    2022

  52. [59]

    S. Milz, F. A. Pollock, and K. Modi, Open Syst. Inf. Dyn. 24, 1740016 (2017)

    2017

  53. [61]

    Baumgratz, M

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

    2014

  54. [62]

    Winter and D

    A. Winter and D. Yang, Phys. Rev. Lett. 116, 120404 (2016)

    2016

  55. [63]

    Yadin, J

    B. Yadin, J. Ma, D. Girolami, M. Gu, and V. Vedral, Phys. Rev. X 6, 041028 (2016)

    2016

  56. [64]

    Kailath and V

    T. Kailath and V. Olshevsky, Integr equ oper theory 22, 65 (1995)

    1995

  57. [65]

    C. Yang, F. C. Binder, M. Gu, and T. J. Elliott, Phys. Rev. E 101, 062137 (2020). 7 Appendix A: Proof of Theorem 1 If two models generate the same process it must hold that ⟨⟨1|EL RR|ρRR⟩⟩ = ⟨⟨1|EL QQ|ρQQ⟩⟩∀L. (A1) Assuming the matrix ERR to be diagonalizable, it can be writte...

    2020

  58. [66]

    Then, αλRQ = 0 = ⇒ αλR=λRQ = αλQ=λRQ

    The eigenvalue λRQ is in both ΛR and ΛQ, i.e, λRQ ∈ ΛR ∩ ΛQ. Then, αλRQ = 0 = ⇒ αλR=λRQ = αλQ=λRQ . (A15) In other words, the coefficients αλ corresponding to the eigenvalue λRQ are the same for both models R and Q

  59. [67]

    Then, αλRQ = 0 (A16) =⇒ αλRQ = αλR=λRQ = 0

    The eigenvalue λRQ is in ΛR but not ΛQ, i.e., λRQ ∈ ΛR but /∈ ΛQ. Then, αλRQ = 0 (A16) =⇒ αλRQ = αλR=λRQ = 0. (A17)

  60. [68]

    𝑖𝑖𝑗𝑗 𝑘𝑘𝑙𝑙

    The eigenvalue λRQ is in ΛQ but not ΛR, i.e., λRQ ∈ ΛQ but /∈ ΛR. Then, αλRQ = 0 (A18) =⇒ αλRQ = αλQ=λRQ = 0. (A19) Therefore, whenever a non-zero eigenvalue appears in only one set (either ΛR or ΛQ), the corresponding coefficients must vanish. Thus we conclude that we must ha...

Pith tools

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