Pith. sign in

REVIEW 3 major objections 9 minor 51 references

Averaging the Rodeo response over random states estimates the density of states as a convolution with a kernel fixed by the evolution-time distribution.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-31 04:34 UTC pith:GOAA7V75

load-bearing objection Clean KPM-style reading of Rodeo that is algebraically solid; the gap is that typicality advantage and preparable states are untested past N=5. the 3 major comments →

arxiv 2607.24972 v1 pith:GOAA7V75 submitted 2026-07-27 quant-ph cond-mat.stat-mech

A Kernel-Based Density of States Estimator for Quantum Computing

Julio Cesar Siqueira Rocha This is my paper
classification quant-ph cond-mat.stat-mech
keywords density of statesRodeo algorithmkernel polynomial methodquantum typicalityspectral kernelHaar-random statestransverse-field Ising modelnear-term quantum algorithms
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The density of states of a quantum many-body system is hard to rebuild once the Hilbert space is too large to diagonalize. This paper shows that the Rodeo algorithm—a simple single-ancilla circuit already used to locate eigenvalues—becomes a direct quantum analogue of the classical kernel polynomial method when its input is a Haar-random state. Averaging the Rodeo signal over such states yields the density of states smoothed by a spectral kernel whose shape is set entirely by how evolution times are sampled. Random states replace stochastic trace estimation; the time distribution replaces the classical damping kernel. Quantum typicality makes the statistical error shrink as the system grows, and the method needs no new circuit hardware. The author derives the estimator and its uncertainties, maps classical window functions onto quantum kernels, and checks the construction on small transverse-field Ising and spin-1 chains.

Core claim

Averaging the Rodeo spectral amplitude over Haar-random input states produces an unbiased estimator of the density of states convolved with a spectral kernel G fixed solely by the temporal sampling distribution p(t). Random states play the role of stochastic trace estimation and p(t) the role of the classical damping kernel, all inside the ordinary single-ancilla Rodeo circuit.

What carries the argument

The spectral kernel G, equal to the characteristic function of the evolution-time distribution p(t) (with a small qudit correction). When the Rodeo response is Haar-averaged, this kernel turns the measured signal into the convolution (g ∗ G)(E), which is the DoS estimator.

Load-bearing premise

That Haar-random or sufficiently design-like input states can actually be prepared at the system sizes where the method is supposed to beat classical enumeration, so that typicality really suppresses the variance.

What would settle it

On a system large enough that exact diagonalization is still possible, prepare approximate random product states, run the Rodeo estimator with a chosen p(t), and check whether the reconstructed DoS (and thermodynamics derived from it) match the exact spectrum within the predicted error bars; a systematic bias that does not shrink with Hilbert-space dimension would falsify the claim.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • Any classical window function (Gaussian, Hann, Hamming, Blackman, Kaiser) becomes a ready-made quantum reconstruction kernel by choosing the matching p(t).
  • Spectral resolution is tuned by the width of p(t); level degeneracies are recovered by integrating reconstructed peaks.
  • Thermodynamic quantities (free energy, mean energy, specific heat) can be read off from the estimated DoS without diagonalizing H.
  • The same single-ancilla circuit already used for eigenvalue location can be reused unchanged for full DoS reconstruction.
  • In the large-system limit the typicality error vanishes, leaving temporal sampling and Trotter error as the dominant uncertainties.

Where Pith is reading between the lines

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

  • Shallow random-product or unitary-design state preparation, already common in classical stochastic trace estimation, is the practical bottleneck that will decide whether the method scales beyond toy models.
  • Pairing the estimator with flat-histogram Monte Carlo (as the author briefly suggests) could extend dynamic range for quantities that need log g(E) over many decades.
  • Hardware noise and decoherence will reshape the effective kernel; characterising that distortion is a natural next experimental target.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 9 minor

Summary. The manuscript shows that the Rodeo eigenvalue-location algorithm, when its response is averaged over Haar-random input states, yields an unbiased estimator of the density of states convolved with a spectral kernel G that is fixed entirely by the characteristic function of the temporal sampling distribution p(t): E[R(E,ψ)] = (g∗G)(E) (Eqs. 11–20). Random states thus play the role of stochastic trace estimation and p(t) the role of the KPM damping kernel, a correspondence the author makes explicit in a "KPM dictionary" (Table I) and a window-function table (Table II). Appendix A derives Var(R) = Σ_i(G_i − Ḡ)²/[N_s(N_s+1)] = O(N_s^{-1}) from the Haar fourth moment, and Appendix B gives a closed-form Hann-kernel expression with |E|^{-3} side-lobe decay. The method is validated against exact diagonalization on the N=5 transverse-field Ising ring (N_s=32) and the N=5 spin-1 ring (N_s=243), including peak-integrated degeneracies (Table III) and thermodynamic quantities (Fig. 5).

Significance. If the construction holds up at scale, it gives a genuinely simple quantum DoS estimator: the standard single-ancilla Rodeo circuit is used unmodified, the reconstruction kernel is parameter-free in the sense of being fixed once p(t) is chosen, and the estimator is provably unbiased with an exactly computed variance (Eq. A12). The explicit dictionary between classical window design and temporal sampling laws (Tables I–II, Appendix B) is a useful methodological contribution that imports decades of signal-processing knowledge into quantum spectral estimation, and the closed-form Hann kernel with its |E|^{-3} leakage floor is a concrete, checkable result. The numerical agreement with exact diagonalization, including the specific heat (which is variance-sensitive), is encouraging. The work is a clean conceptual unification rather than a demonstrated computational advantage: the central algebraic claims are sound, but the practical case for the regime where the method is "aimed" (Hilbert spaces beyond classical enumeration) rests on assumptions about state preparation and evolution-time cost that the manuscript does not yet quantify.

major comments (3)
  1. [Sec. II.C, Appendix A, Sec. IV (limitations paragraph)] The typicality scaling σ(ĝ)=O(N_s^{-1/2}) (Eq. 21, Eq. A14) is derived from exact Haar second and fourth moments (Eq. A3), yet the paper's stated target is the regime where exact Haar states are exponentially costly to prepare, and the Conclusion asserts that 'a (pseudo)random product of local rotations' will suffice, citing classical KPM practice. This citation does not settle the question: classical KPM random vectors have independent entries (Gaussian or ±1), which is not the same ensemble as shallow-circuit or random-product states on a quantum register. Two things can be stated precisely and would substantially strengthen the paper. (i) Unbiasedness requires only E[|ψ⟩⟨ψ|] = I/N_s, which holds exactly for random product states with local Haar-random (or even discrete 1-design) single-site rotations — so the estimator's mean is safe under far weaker assumptions than the text suggests
  2. [Secs. II.B, II.D.2, III.B (Fig. 3), Eq. (27)] There is an unresolved internal tension between spectral resolution and simulation cost that bears directly on the method's viability. Narrowing the Gaussian kernel requires large σ (Eq. 16), and the sampling law p(t)=N(0,σ) has unbounded support, so typical draws have |t|~σ and tails to several σ. With the first-order Trotter choice r=t²/δ (Eq. 27), the σ=200 refinement in Fig. 3 implies draws with t~200–600, i.e. r up to ~10^6 Trotter steps per controlled evolution at δ=0.05 — many orders of magnitude beyond any near-term hardware, and the dominant cost of the whole protocol. The paper acknowledges the trade-off qualitatively but never aggregates it: there is no estimate of total evolution time (or total controlled-U count) per reconstructed DoS point, and no comparison of that cost against the sparse matvec count of classical KPM for the same resolution. Since the abstract and Conclus
  3. [Sec. III (Figs. 2–5, Table III)] All validation is at N_s=32 and N_s=243, where exact diagonalization is trivial and where the O(N_s^{-1/2}) typicality asymptotics say nothing — with R=10–36 states the observed errors are dominated by the R^{-1/2} factor, not by the dimension suppression that is the paper's headline mechanism. This is acknowledged only obliquely. The paper would be considerably more convincing with one scaling demonstration: a sequence of system sizes (e.g. N=4,6,8,10 for the spin-1/2 TFIM, still classically checkable) showing the predicted N_s^{-1/2} decay of the single-state variance and the quality of the R-averaged estimator as N_s grows. This is cheap classically, directly tests Eq. A12 rather than merely illustrating the estimator, and would substantiate the claim that the method improves precisely where classical enumeration becomes prohibitive.
minor comments (9)
  1. [Sec. II.A, Eq. (6)] The definition 'N_s = d^N_s' (rendered as N s =d N s) is typographically ambiguous; please write N_s = d_s^N explicitly.
  2. [Sec. II.D.1, Eq. (23)] The symbol σ is overloaded: it is the width of the Gaussian sampling law (Eq. 13) and is reused in Eq. (23) for the total standard deviation of the estimator. Please use a distinct symbol for the latter.
  3. [Sec. II.D.2, Eq. (27)] r = t²/δ is presented as a definition ('we consider the time dependence...') rather than a consequence of the first-order Trotter error bound; please state the assumption (norm of the commutator sum absorbed into δ, or cite the standard scaling, e.g. Ref. [27]) and justify the choice δ=0.05 used throughout Sec. III.
  4. [Table III] Entries such as ĝ=0.01(1) imply a ~100% relative uncertainty, yet a relative deviation ∆rg=0.26 is quoted to two digits without comment. Please clarify how the uncertainties in parentheses were propagated and whether the ∆rg values are meaningful given those error bars.
  5. [Fig. 1 vs. Eq. (3)] The phase-shift gate is labeled P(ϕ) in the figure but defined as P(E,t) in Eq. (3); please harmonize the notation.
  6. [Sec. II.B, Eqs. (13)–(14)] The Gaussian is introduced with mean µ which is then set to zero; consider defining the zero-mean law directly and mentioning the µ≠0 phase issue in one sentence.
  7. [Sec. III header] Typo: 'RESUL TS' in the section heading.
  8. [Data Availability] The repository is listed as 'in preparation'. Given that the numerical comparisons are a central part of the validation, the code and raw data should be deposited (with a citable DOI) before publication.
  9. [Sec. III.C, Fig. 4] The 0.1-width binning of exact eigenvalues is described as 'for visualization purposes only', but the red reference curve/markers in Fig. 4 should be explicitly identified as binned in the caption so readers do not mistake bin height for degeneracy.

Circularity Check

1 steps flagged

No significant circularity: the Haar-average DoS estimator is derived algebraically in-manuscript and validated against exact diagonalization; self-citations supply only the prior Rodeo circuit, which is re-derived here.

specific steps
  1. self citation load bearing [Sec. II.A / Refs. [20, 21]; also Sec. I]
    "we employ the qudit implementation of the Rodeo algorithm introduced in Ref. [21]. The complete circuit construction is described therein; here we summarize only the ingredients required to derive the DoS estimator."

    The single-shot Rodeo kernel K_da is taken from the author’s prior work. This is not strongly circular: the manuscript re-derives the expectation value (Eqs. 7–8) from the circuit, and the novel step (Haar average → g ∗ G) does not rely on an unverified uniqueness claim in [20, 21]. Flagged only as minor self-citation burden, not as a reduction of the central claim.

full rationale

The central claim—that the Haar average of the Rodeo spectral amplitude equals the DoS convolved with a kernel G fixed by p(t)—is obtained by direct calculation in Secs. II.A–II.C (Eqs. 7–20) and Appendix A, using the Haar second/fourth moments and the characteristic function of p(t). Nothing in that chain is fitted to the target DoS, nor is any uniqueness theorem imported to force the result. Numerical checks compare the estimator to exact diagonalization of the same small Hamiltonians (N_s = 32 and 243), an external benchmark. Self-citations [20, 21] point to the author’s earlier Rodeo-kernel papers; those supply context and the single-shot response formula, but the present manuscript re-derives the ancilla expectation (Eqs. 7–8) from the circuit and then performs the new Haar-average and KPM-dictionary steps independently. That is ordinary cumulative self-citation, not a load-bearing circular reduction. Score 1 reflects only that minor dependence; the derivation does not collapse to its inputs by construction.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

The claim rests on standard Haar moment formulas, the already-published Rodeo response kernel, unitary time evolution (possibly Trotterized), and the modeling choice that p(t) fully specifies the reconstruction kernel. No new physical entities are postulated. Free parameters are algorithmic knobs (σ, R, N_t, δ, da) chosen for numerics, not fitted to hide discrepancy with exact DoS.

free parameters (4)
  • Gaussian sampling width σ = 20 or 200 (dimensionless time)
    Sets spectral kernel width and thus resolution vs. evolution-time cost; chosen as σ=20 (main figures) or σ=200 (refinement), not derived from a principle.
  • Number of Haar states R and evolution samples N_t = N_t=1000; R=1,10, or 36
    Finite-sample controls on estimator variance; set to R=1–36 and N_t=1000 for reported runs.
  • Trotter target precision δ = 0.05
    Hand-chosen accuracy for first-order Suzuki decomposition via r=t²/δ; affects systematic bias not folded into plotted error bars.
  • Ancilla dimension d_a = 3
    Algorithmic choice (qutrit used throughout) that reshapes the two-component spectral kernel; not a physical constant.
axioms (5)
  • standard math Haar-random pure states have E[|c_k|^2]=1/N_s and fourth moments E[|c_i|^2|c_j|^2]=(1+δ_ij)/(N_s(N_s+1))
    Used in Sec. II.C and Appendix A to obtain unbiasedness and typicality variance scaling.
  • domain assumption Single-shot Rodeo ancilla expectation for fixed t equals ∑_k |c_k|^2 K_{d_a}(E_k−E,t) with kernel as in Eq. 8
    Taken from the qudit Rodeo construction (Ref. [21]); this paper averages that response rather than re-deriving the circuit.
  • domain assumption Time evolution e^{-iHt} is accurately realized (exactly or via Suzuki–Trotter at chosen δ)
    Sec. II.D.2; required for the measured signal to equal the ideal spectral amplitude.
  • standard math Off-diagonal coherences do not contribute because the measured observable is Z_{d_a}⊗1_s
    Sec. II.A; justifies dependence only on populations |c_k|^2.
  • ad hoc to paper Approximate unitary designs or random product states will inherit enough typicality for the estimator to remain useful at large N_s
    Stated as a hoped extension in Sec. IV; not demonstrated, yet required for the method's intended large-system regime.

pith-pipeline@v1.2.0-grok45-kimik3 · 20418 in / 3633 out tokens · 69198 ms · 2026-07-31T04:34:37.542083+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of A Kernel-Based Density of States Estimator for Quantum Computing." pith.science (2026). https://pith.science/paper/GOAA7V75

@misc{pith2026260724972,
  author       = {Pith},
  title        = {Pith review of: A Kernel-Based Density of States Estimator for Quantum Computing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GOAA7V75}},
  note         = {Machine review of arXiv:2607.24972}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

The density of states (DoS) encodes the thermodynamic and spectral properties of quantum many-body systems, yet its reconstruction becomes intractable for Hilbert spaces too large to diagonalize. Classically, the kernel polynomial method (KPM) addresses this by combining stochastic trace estimation with a smoothing kernel. Here we show that the Rodeo algorithm---one of the simplest eigenvalue-location protocols for near-term quantum hardware---provides a direct quantum analogue of this approach. Averaging the Rodeo response over Haar-random input states yields the DoS convolved with a spectral kernel fixed entirely by the distribution of evolution times: the random states play the role of stochastic trace estimation, and the temporal sampling distribution that of the damping kernel. The construction requires only the standard single-ancilla circuit, and quantum typicality suppresses the statistical error as the Hilbert-space dimension grows. We derive the estimator and its uncertainties, establish an explicit dictionary between signal-processing window functions and quantum reconstruction kernels, and validate the method on the one-dimensional transverse-field Ising and spin-1 models.

Figures

Figures reproduced from arXiv: 2607.24972 by Julio Cesar Siqueira Rocha.

Figure 1
Figure 1. Figure 1: FIG. 1. Quantum circuit implementing a single Rodeo cycle with an ancilla qudit of dimension [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. DoS reconstruction for the 1D transverse-field Ising model using the Rodeo kernel method. For each target energy, the [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Refinement of the DoS reconstruction for the previously merged peaks shown in Fig. 2(b) for the transverse-field [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. DoS reconstruction for the one-dimensional spin-1 model using the Rodeo kernel method. For each target energy, the [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Thermodynamic properties of the spin-1 ring with [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

51 extracted references · 2 canonical work pages

  1. [1]

    Statistical uncertainties The first source of statistical uncertainty originates from the use of Haar-random input states. Owing to quantum typicality [22, 23], the spectral weights|c k|2 follow a symmetric Dirichlet distribution whose fluctua- tions are suppressed by the Hilbert-space dimension (see Appendix A). As a consequence, the standard deviation o...

  2. [2]

    The most important of these is the Suzuki–Trotter decomposition employed to simulate the time evolution generated by Hamiltonians containing noncommuting terms

    Implementation errors Besides the statistical uncertainties discussed above, the quantum implementation introduces additional sys- tematic errors. The most important of these is the Suzuki–Trotter decomposition employed to simulate the time evolution generated by Hamiltonians containing noncommuting terms. Let the Hamiltonian be decomposed as H= KX k=1 Hk...

  3. [3]

    Expanding the square, R2 = X ij |ci|2|cj|2GiGj.(A2) 11 whereG i =G(E−E i), withE i being thei-th energy eigenvalue

    V ariance The variance follows from Var(R) =E(R 2)−E(R) 2,(A1) where we adoptedR=R da (E, ψ) to simplify the nota- tion. Expanding the square, R2 = X ij |ci|2|cj|2GiGj.(A2) 11 whereG i =G(E−E i), withE i being thei-th energy eigenvalue. The only ingredient required is the Haar fourth mo- ment, E cic∗ j ckc∗ l = δijδkl +δ ilδjk Ns(Ns + 1) ,(A3) which immed...

  4. [4]

    Typicality scaling Since the filterG(∆) is bounded independently of the Hilbert-space dimension, X i (Gi − G)2 =O(N s),(A13) and therefore Var(R) =O(N −1 s ),(A14) or equivalently, σ(R) =O(N −1/2 s ).(A15) Thus the fluctuations of the Rodeo response decrease as the inverse square root of the Hilbert-space dimension, which is the characteristic signature o...

  5. [5]

    V ariance of the DoS estimator For the estimator ˆg(E) =1 R RX r=1 Rda (E, ψr),(A16) constructed fromRindependent Haar-random states, the sample mean is unbiased,E(ˆg) =E(R), and its variance is Var(ˆg) =1 R Var(R).(A17) Consequently, at fixedR, σ(ˆg) =O(N−1/2 s ).(A18) SinceN s =d N s , the variance decreases exponentially with the system size and conseq...

  6. [6]

    Goold, M

    J. Goold, M. Huber, A. Riera, L. d. Rio, and P. Skrzypczyk, Journal of Physics A: Mathematical and Theoretical49, 143001 (2016)

    2016

  7. [7]

    D. H. E. Gross,Microcanonical Thermodynamics: Phase Transitions in “Small” Systems, Lecture Notes in Physics, Vol. 66 (World Scientific, Singapore, 2001). 13

    2001

  8. [8]

    J. C. S. Rocha, R. A. Dias, and B. V. Costa, Phys. Rev. E 112, 014112 (2025), https://arxiv.org/abs/2502.00999

    arXiv 2025

  9. [9]

    R. Kubo, J. Phys. Soc. Jpn.12, 570 (1957)

    1957

  10. [10]

    G. D. Mahan,Many-Particle Physics, 3rd ed. (Kluwer Academic/Plenum Publishers, New York, 2000)

    2000

  11. [11]

    N. W. Ashcroft and N. D. Mermin,Solid State Physics (Holt, Rinehart and Winston, New York, 1976)

    1976

  12. [12]

    Qi and M

    K. Qi and M. Bachmann, Phys. Rev. Lett.120, 180601 (2018)

    2018

  13. [13]

    Weiße, G

    A. Weiße, G. Wellein, A. Alvermann, and H. Fehske, Rev. Mod. Phys.78, 275 (2006)

    2006

  14. [14]

    R. N. Silver and H. R¨ oder, Int. J. Mod. Phys. C5, 735 (1994)

    1994

  15. [15]

    R. P. Feynman, Int. J. Theor. Phys.21, 467 (1982)

    1982

  16. [16]

    Lloyd, Science273, 1073 (1996)

    S. Lloyd, Science273, 1073 (1996)

    1996

  17. [17]

    D. S. Abrams and S. Lloyd, Phys. Rev. Lett.83, 5162 (1999)

    1999

  18. [18]

    Roggero, Phys

    A. Roggero, Phys. Rev. A102, 022409 (2020)

    2020

  19. [19]

    R. D. Somma, New J. Phys.21, 123025 (2019)

    2019

  20. [20]

    T. E. O’Brien, B. Tarasinski, and B. M. Terhal, New J. Phys.21, 023022 (2019)

    2019

  21. [21]

    Lin and Y

    L. Lin and Y. Tong, PRX Quantum3, 010318 (2022)

    2022

  22. [22]

    K. Choi, D. Lee, J. Bonitati, Z. Qian, and J. Watkins, Phys. Rev. Lett.127, 040505 (2021)

    2021

  23. [23]

    Z. Qian, J. Watkins, G. Given, J. Bonitati, K. Choi, and D. Lee, Eur. Phys. J. A60, 151 (2024)

    2024

  24. [24]

    R. F. I. Gomes, J. C. S. Rocha, W. A. T. Nogueira, and R. A. Dias, Physica Scripta100, 065119 (2025), https://arxiv.org/abs/2407.11301

    arXiv 2025

  25. [25]

    J. C. S. Rocha, R. F. I. Gomes, W. A. T. Nogueira, and R. A. Dias, Quantum Information Processing23, 10.1007/s11128-024-04552-1 (2024), https://arxiv.org/abs/2312.04322

    Pith/arXiv arXiv 2024

  26. [26]

    J. C. S. Rocha and R. A. Dias (2026), https://doi.org/10.48550/arXiv.2603.16049

    work page doi:10.48550/arxiv.2603.16049 2026

  27. [27]

    Goldstein, J

    S. Goldstein, J. L. Lebowitz, R. Tumulka, and N. Zangh ` ı, Phys. Rev. Lett.96, 050403 (2006)

    2006

  28. [28]

    Popescu, A

    S. Popescu, A. J. Short, and A. Winter, Nat. Phys.2, 754 (2006)

    2006

  29. [29]

    ˙Zyczkowski and H.-J

    K. ˙Zyczkowski and H.-J. Sommers, J. Phys. A: Math. Gen.34, 7111 (2001)

    2001

  30. [30]

    Suzuki, Journal of Mathematical Physics32, 400 (1991), https://pubs.aip.org/aip/jmp/article- pdf/32/2/400/8160505/400 1 online.pdf

    M. Suzuki, Journal of Mathematical Physics32, 400 (1991), https://pubs.aip.org/aip/jmp/article- pdf/32/2/400/8160505/400 1 online.pdf

    1991

  31. [31]

    Hatano and M

    N. Hatano and M. Suzuki, Finding exponential prod- uct formulas of higher orders, inQuantum Annealing and Other Optimization Methods, edited by A. Das and B. K. Chakrabarti (Springer Berlin Heidelberg, Berlin, Heidelberg, 2005) pp. 37–68

    2005

  32. [32]

    A. M. Childs, Y. Su, M. C. Tran, N. Wiebe, and S. Zhu, Phys. Rev. X11, 011020 (2021)

    2021

  33. [33]

    Barenco, C

    A. Barenco, C. H. Bennett, R. Cleve, D. P. DiVincenzo, N. Margolus, P. Shor, T. Sleator, J. A. Smolin, and H. Weinfurter, Phys. Rev. A52, 3457 (1995)

    1995

  34. [34]

    Bernstein and U

    E. Bernstein and U. Vazirani, SIAM Journal on Comput- ing26, 1411 (1997)

    1997

  35. [35]

    Bouland, B

    A. Bouland, B. Fefferman, C. Nirkhe, and U. Vazirani, Nature Phys.15, 159 (2018), arXiv:1803.04402 [quant- ph]

    Pith/arXiv arXiv 2018

  36. [36]

    Nielsen and I

    M. Nielsen and I. Chuang, Quantum computation and quantum information: 10th anniversary edition (Cam- bridge University Press, 2010) Chap. 7

    2010

  37. [37]

    Jaeger,Quantum Information: An Overview (Springer New York, 2006)

    G. Jaeger,Quantum Information: An Overview (Springer New York, 2006)

    2006

  38. [38]

    S. J. Devitt, W. J. Munro, and K. Nemoto, Reports on Progress in Physics76, 076001 (2013)

    2013

  39. [39]

    Joint Committee for Guides in Metrology, Evaluation of measurement data – guide to the expression of uncer- tainty in measurement (2008), jCGM 100:2008

    2008

  40. [40]

    Sachdev,Quantum Phase Transitions, 2nd ed

    S. Sachdev,Quantum Phase Transitions, 2nd ed. (Cam- bridge University Press, Cambridge, 2011)

    2011

  41. [41]

    Lucas, Front

    A. Lucas, Front. Phys.2, 5 (2014)

    2014

  42. [42]

    Kochenberger, J.-K

    G. Kochenberger, J.-K. Hao, F. Glover, M. Lewis, Z. L¨ u, H. Wang, and Y. Wang, J. Comb. Optim.28, 58 (2014)

    2014

  43. [43]

    Kadowaki and H

    T. Kadowaki and H. Nishimori, Phys. Rev. E58, 5355 (1998)

    1998

  44. [44]

    Farhi, J

    E. Farhi, J. Goldstone, S. Gutmann, J. Lapan, A. Lund- gren, and D. Preda, Science292, 472 (2001)

    2001

  45. [45]

    Albash and D

    T. Albash and D. A. Lidar, Rev. Mod. Phys.90, 015002 (2018)

    2018

  46. [46]

    Hauke, H

    P. Hauke, H. G. Katzgraber, W. Lechner, H. Nishimori, and W. D. Oliver, Rep. Prog. Phys.83, 054401 (2020)

    2020

  47. [47]

    Zeeman, The London, Edinburgh, and Dublin Philo- sophical Magazine and Journal of Science43, 226 (1897), https://doi.org/10.1080/14786449708620985

    P. Zeeman, The London, Edinburgh, and Dublin Philo- sophical Magazine and Journal of Science43, 226 (1897), https://doi.org/10.1080/14786449708620985

    work page doi:10.1080/14786449708620985

  48. [48]

    D. J. Griffiths and D. F. Schroeter, Introduction to quan- tum mechanics (Cambridge University Press, Cambridge ; New York, NY, 2018) Chap. 7, pp. 389–416, third edi- tion ed

    2018

  49. [49]

    Pfeuty, Annals of Physics57, 79 (1970)

    P. Pfeuty, Annals of Physics57, 79 (1970)

    1970

  50. [50]

    Wang and D

    F. Wang and D. P. Landau, Phys. Rev. Lett.86, 2050 (2001)

    2050

  51. [51]

    B. A. Berg and T. Neuhaus, Phys. Rev. Lett.68, 9 (1992)

    1992