Pith. sign in

REVIEW 4 major objections 6 minor 43 references

Canonical Partition Function on a Quantum Computer through Trotter Interpolation

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A quantum algorithm estimates the canonical partition function through Trotterized evolution and Chebyshev interpolation, with cost quasi-linear in inverse temperature and logarithmic in error, using 2n+2 qubits.

desk verdict Promising idea, but a missing e^{-β} factor in the amplitude estimation invalidates the headline cost, so I'd reject as submitted. read the letter →

arxiv 2506.09318 v1 pith:TSVMN3D5 submitted 2025-06-11 quant-ph

Taozhi Guo , Gumaro Rendon , Rutuja Kshirsagar This is my paper · ORCID
classification quant-ph
keywords canonicalpartitionfunctionGibbsstateTrotterizationChebyshevinterpolationgeneralizedquantumsignalprocessingSYKmodelthermalobservablesancillaqubits
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 that the thermal partition function Z(β)=Tr[$e^{{-βH}}$] of a many-qubit Hamiltonian can be estimated without block-encoding the Hamiltonian, by instead applying generalized quantum signal processing to the Trotterized time-evolution operator and then interpolating the results to zero Trotter step size. Concretely, the authors claim a quantum computational cost of $\tilde O(\beta \max(\alpha,1)\log(1/\epsilon)/(\epsilon \sqrt{\max_s Z(\beta,s)/N}))$ for statistical error $\epsilon$ and algorithmic error $\epsilon$, with only $2n+2$ qubits. This would match the scaling of qubitization-based thermal-state algorithms while saving $\log \Gamma$ ancilla qubits when the Hamiltonian has $\Gamma$ fast-forwardable fragments. This matters because the Trotterized operator is easier to implement than a quantum walk or block-encoding, and the qubit saving grows with the number of Hamiltonian terms, as the SYK model benchmark illustrates.

What carries the argument

The machinery has three pieces. The effective Hamiltonian $\tilde H_p(\tau)=\log S_p(\tau)/(i\tau)$ converts a $p$-th order Trotter product formula $S_p(\tau)$ into an exactly implemented but approximate Hamiltonian; the paper then implements $e^{-\beta(\tilde H_p+1)/2}$ by generalized quantum signal processing applied to powers of $S_p^{\lceil 1/(s t)\rceil}(s t)$. Chebyshev interpolation in the Trotter step variable $s$ transports the measured traces from the finite-step nodes to $s=0$, where the effective Hamiltonian coincides with $H$. The cost estimate flows from the low-weight Fourier expansion degree $M_k=O(\beta \log(1/\epsilon_{\mathrm{QSP}}))$, the Chebyshev interpolation order $M_{\mathrm{cheb}}=\tilde O(\log(Z/N\epsilon_{\mathrm{cheb}})+\alpha)$, and amplitude estimation on the thermofield-double state to read out each trace.

What would settle it

For a small exactly diagonalizable Hamiltonian (or the SYK model with n=8 Majoranas), compute Z(β,s_k)/N exactly at the Chebyshev nodes, build the interpolated value with the paper's d_k coefficients, and compare with the exact Z(β)/N across a range of β; if the required M_cheb grows faster than \tilde O(\log(1/ε)+α) or the error decays slower than $ρ^{{-M_cheb}}$, the central claim is falsified. A direct check is to evaluate max_{z∈B_ρ} |Tr[$e^{{-β \tilde H_p(z t)}}$]|/N numerically and test the bound of Eq. (41).

Watch

Extended reading notes

Core claim

The central claim, stated as Theorem V.4, is that for $H=\sum_\gamma H_\gamma$ with fast-forwardable $H_\gamma$ and $\|H\|\le 1$, one can estimate $Z(\beta)/N$ by estimating the traces $Z(\beta,s_k)/N = \operatorname{Tr}[e^{-\beta \tilde H_p(s_k t)}]/N$ at Chebyshev nodes $s_k$, forming the linear combination with coefficients $d_k$, and reading the result as the $s\to 0$ limit. The extrapolation is justified by bounding the Chebyshev interpolation error of the analytic function $f(s)=Z(\beta,s)/N$ on a Bernstein ellipse, using a Trotter-type bound on the difference between the effective Hamiltonian $\tilde H_p(\tau)=\log S_p(\tau)/(i\tau)$ and $H$. The claimed cost is $\tilde O(\beta \max(\alpha,1)\log(1/\epsilon)/(\epsilon \sqrt{\max_s Z(\beta,s)/N}))$ using $2n+2$ qubits, which the paper presents as matching qubitization-based methods while avoiding the $\log \Gamma$ ancilla overhead.

Load-bearing premise

The whole result rests on the assumption that the trace of the Trotterized thermal operator remains well-behaved, with errors growing only mildly, when the Trotter step size is allowed to be complex; if that analytic control fails, the claimed logarithmic error scaling and the headline cost do not follow.

Editorial extensions

If this is right

  • For any Hamiltonian satisfying the fast-forwardable-fragment assumption, the partition function can be estimated in $\tilde O(\beta \log(1/\epsilon))$ quantum operations up to the $1/(\epsilon \sqrt{\max_s Z(\beta,s)/N})$ overhead, so the inverse-temperature dependence is effectively quasi-linear.
  • The algorithm uses $2n+2$ qubits, saving $\log \Gamma$ ancilla qubits relative to qubitization-based methods; for Hamiltonians with many fragments, such as the SYK model with $O(n^4)$ terms, this saving grows with system size.
  • Because the trace rather than the spectral norm controls the interpolation error, the algorithmic error enters as $\log(1/\epsilon)$ instead of $\log^2(1/\epsilon)$.
  • The method replaces the quantum walk operator with the Trotterized evolution operator, avoiding the controlled operations needed for block-encoding.
  • Access to the trace estimates also gives, through the thermofield-double construction, a route to thermal expectation values, not just to the partition function itself.

Reading between the lines

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

  • The paper stops at the partition function; one could extend the same interpolation to thermal expectation values of observables by weighting the thermofield-double amplitude estimates with the observable, which would make the method directly useful for phase-transition studies.
  • The denominator $\sqrt{\max_s Z(\beta,s)/N}$ implies the practical cost worsens at low temperature; a small-system exact-diagonalization comparison could reveal whether this factor, not the $\log(1/\epsilon)$ term, dominates in realistic regimes.
  • The central risk is the bound behind Eq. (41), which is imported from another work; a standalone proof or direct numerical check of that bound for representative models would turn the cost claim from conditional into demonstrated.
  • Because the ancilla saving grows as $\log \Gamma$, the method should become most attractive for Hamiltonians with many commuting fragments, such as molecular Hamiltonians after fermion-to-qubit mappings, a regime the paper does not test.
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 / 6 minor

Summary. The paper proposes a quantum algorithm to estimate the canonical partition function Z(β)/N for a Hamiltonian H = Σ_γ H_γ with fast-forwardable terms, using Trotter-Suzuki product formulas, generalized quantum signal processing (GQSP), and Chebyshev interpolation in the Trotter step size. The central claim is a quantum computational cost of Õ(β log(1/ε)/(ε√(max_s Z(β,s)/N))) using 2n+2 qubits, achieved by evaluating the trace of the Boltzmann factor for a Trotterized effective Hamiltonian at several Trotter step sizes and interpolating to zero step size. The authors also present numerical results for the SYK model, focusing on the polynomial approximation error and ancilla-qubit savings relative to qubitization.

Significance. If the central claim were correct, the method would match the asymptotic scaling of qubitization-based Gibbs-state algorithms while saving O(log Γ) ancilla qubits, which is a useful practical improvement. The use of Chebyshev interpolation to remove Trotter error is an interesting idea, and the paper provides explicit cost estimates and a numerical check of the GQSP polynomial approximation. However, the main cost claim is invalidated by an internal inconsistency in the amplitude-estimation analysis: the probability computed from the defined operator is not the quantity used in the cost derivation, and the correction changes the query complexity by an exponential factor in β. Because this affects the central theorem, the paper cannot be accepted in its current form.

major comments (4)
  1. [Section V.A, Eq. (25) and Eq. (28)] The operator U_boltz,s defined in Eq. (25) contains the shift e^{-β(\tilde H_p+1)/2}, so the probability of measuring |0> on the ancilla is p0 = (1/N) Σ_n e^{-β(\tilde λ_n+1)} = e^{-β} Z(β,s)/N. Equation (28) instead states p0 = Z(β,s)/N, dropping the e^{-β} factor. This is not a typographical nuisance: to estimate q = Z(β,s)/N with additive error ε, the amplitude a0 = √p0 must be estimated with error δ_a = e^{-β/2} ε/(2√q), so the number of U_boltz queries in Lemma V.2 and Theorem V.4 becomes O(e^{β/2} √q / ε) rather than O(√q / ε). The claimed cost Õ(β log(1/ε)/(ε√(max_s Z(β,s)/N))) therefore does not follow, and the error is internal to the manuscript rather than a matter of outside consensus.
  2. [Section II and Lemma V.1, Eq. (26)] The algorithm statement in Section II defines Z(β,s_k) = Tr[e^{-iβ \tilde H_p(s_k t)}] with an imaginary unit, whereas Lemma V.1 and Eq. (26) use the real Boltzmann factor e^{-β(\tilde H_p+1)/2}. These are incompatible definitions of the quantity to be estimated. In addition, the definition of β_k in the proof of Lemma V.1 is ambiguous as written; if β_k = β ⌈1/(s_k t)⌉ / (1/(s_k t)), then the exponent in Eq. (26) becomes -β (m s_k t)^2 \tilde H_p with m = ⌈1/(s_k t)⌉, which does not give the intended inverse temperature β. The correction factor needs to be stated unambiguously and verified.
  3. [Section V.B, Eq. (41) and Ref. [33]] The Chebyshev interpolation error analysis relies on the bound max_{z∈B_ρ} |Tr[e^{-β \tilde H_p(z t)}]/N| = O(e^{β(r t)^p α/p!} Z/N), which is asserted to follow from Lemma 12 of Ref. [33], an unpublished arXiv preprint. This analytic-continuation statement is load-bearing: without it, the bound on ε_cheb and hence the scaling M_cheb = Õ(log(Z/N)/ε_cheb + α) in Lemma V.3 do not hold. The manuscript should provide a self-contained proof of this bound, or at minimum a careful statement of the analyticity domain of \tilde H_p(τ) for complex τ and the justification for replacing the disc of radius r with the Bernstein ellipse B_ρ.
  4. [Section VI] The numerical section is presented as a benchmark against state-of-the-art methodology using the SYK model, but the simulations only test the polynomial approximation of e^{-β(x+1)} (Taylor vs. LWF) and the number of ancilla qubits saved. There is no numerical demonstration of the full algorithm: no computation of Z(β,s_k) via amplitude estimation, no Chebyshev interpolation to s=0, and no comparison with a qubitization-based estimate. The statement that the results 'fit the analytical derivation' is therefore an overclaim with respect to the paper's central algorithm.
minor comments (6)
  1. [Abstract and Theorem V.4] The abstract quotes the cost as Õ(β log 1/ε), which omits the 1/ε and 1/√(max_s Z/N) factors that appear in Theorem V.4. Please quote the full scaling to avoid misleading readers.
  2. [Section II] The bullet point 'we construct an implementation of the operator e^{-iβ \tilde H_p(s t)}' should read e^{-β \tilde H_p(s t)} without the imaginary unit, consistent with the rest of the paper.
  3. [Section V.A, Eq. (28)] Equation (28) should be p0 = e^{-β} Z(β,s)/N, reflecting the +1 shift in Eq. (25). This is a direct consequence of the operator definition and should be corrected throughout the cost analysis.
  4. [Throughout] There are several typographical errors, including 'Majorana Majorana Fermions' in the Introduction, 'Futher' near the end of the Introduction, 'a with statistical error' in Theorem V.4, and the reference to 'Figure. 1' in Section VI as a simulation result when Figure 1 is actually the circuit diagram.
  5. [Eq. (44) and Lemma V.3] The symbol M_cheb is used both for the number of Chebyshev nodes and for the degree of the interpolant; the paper should distinguish these notions or state explicitly that they are related by M_cheb − 1.
  6. [References] Ref. [33] is an unpublished arXiv preprint. Because the paper relies on it for a central bound in Section V.B, either a published version should be cited or the proof should be included in an appendix.

Circularity Check

1 steps flagged · score 4.0 of 10

Central Trotter-interpolation bound is imported from a same-author unpublished preprint; the rest of the algorithm is not definitionally circular, though an internal e^{-β} factor error affects the cost claim.

  1. self citation load bearing [Section V.B, Eq. (35), feeding Eqs. (40)-(41), Lemma V.3 and Theorem V.4]
    "Starting from the Trotterization error, from Lemma 12 in Ref. [33], the difference between the Hamiltonian and the effect Hamiltonian scales as ||L_p(τ)||=||H̃_p(τ)−H||=O(|τ|^p α/(p+1)!)."

    The analytic-continuation bound max_{z∈B_ρ}|f(z)|=O(e^{β(rt)^p α/p!} Z/N) (Eq. 41), which controls the Chebyshev interpolation error and hence M_cheb in Lemma V.3 and the final \tilde{O}(β log 1/ε) cost in Theorem V.4, is built directly on Lemma 12 of Ref. [33]. Ref. [33] is an unrefereed arXiv preprint by one of the present authors (G. Rendon, arXiv:2311.01533); it is not proved in this paper and is not supported by a machine-checked or code-reproduced verification. The headline cost therefore rests, at its decisive Trotter-interpolation step, on a same-author citation rather than on a self-contained derivation.

full rationale

The algorithm is not circular in the fitted-parameter sense: polynomial orders M_k and M_cheb are chosen from error targets, and the SYK numerics are an independent benchmark. The Chebyshev interpolation at s=0 is a genuine extrapolation from estimated values Z(β,s_k)/N, not a reuse of the target. The main circularity signal is load-bearing self-citation: the crucial Trotter-error bound Eq. (35) and the Bernstein-ellipse bound Eq. (41) are taken from Lemma 12 of Ref. [33], an unrefereed same-author preprint, so the log(1/ε) interpolation error and Theorem V.4's cost are not self-contained. A secondary self-citation ([32], used for the Chebyshev-node sum and interpolation prescription) is standard material and less concerning. I note also that Eq. (28) drops the e^{-β} factor from the +1 shift in Eq. (25), so the measured probability is e^{-β}Z(β,s)/N rather than Z(β,s)/N; this is an internal scaling error that changes the amplitude-estimation query count, but it is a correctness issue rather than a circularity, and I do not count it in the score. Overall: some self-citation, with the central claim still containing independent algorithmic content, so 4.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central result rests on two nonstandard assumptions: (1) the Hamiltonian decomposition satisfies Σ_γ||H_γ||≤1, and (2) the effective Trotter Hamiltonian error bound of Ref. [33] extends to complex step sizes within a Bernstein ellipse. The latter is a self-cited, unrefereed lemma. The free parameters t, p, r, and δ are analytic choices, not data fits.

free parameters (4)
  • Trotter step scale t = e^{-sqrt(log β) log 5}
    Chosen in Eq. (45)/(58) to balance Trotter error against circuit depth; not fitted to data.
  • Product formula order p = sqrt(log^5 β) in Eq. (45), but Eq. (59) requires sqrt(log β)
    The paper's stated scaling is inconsistent; the cost formula depends on this choice.
  • Chebyshev ellipse radius r = e^{1/sqrt(log^5 β)}
    Chosen to bound the analytic continuation region; enters M_cheb via Eq. (44).
  • Shift δ = ~1/β
    Introduced in Section III.A to keep the argument of e^{-β(x+1)} away from the singular point x=-1.
assumptions (5)
  • domain assumption Each H_γ is fast-forwardable and Σ_γ ||H_γ|| ≤ 1
    Assumed at the top of Section II; the Trotter error bound α and the GQSP degree depend on it. Fails for generic local Hamiltonians without renormalization.
  • standard math GQSP theorem and angle-finding algorithm of Ref. [29]
    Quoted in Theorem IV.1; used to implement the polynomial transformation P(S_p^N) with degree d.
  • standard math Chebyshev interpolation error bound for analytic functions on Bernstein ellipse (Theorem 8.2 of Ref. [35])
    Used as Lemma III.2 for the interpolation error.
  • domain assumption Effective Hamiltonian error bound ||\tilde{H}_p(τ)-H||=O(|τ|^p α/(p+1)!) for complex τ (Lemma 12 of Ref. [33])
    Central to the interpolation error estimate; cited from the authors' unpublished arXiv preprint, not re-derived in this paper.
  • ad hoc to paper Analytic continuation of f(s)=Tr[e^{-β\tilde{H}_p(s t)}]/N to the Bernstein ellipse B_ρ with the bound in Eq. (41)
    The log branch of S_p(τ) for complex τ and the norm growth e^{|Im τ| Σ ||H_γ||} are not rigorously controlled; this is where the tight log(1/ε) scaling enters.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Canonical Partition Function on a Quantum Computer through Trotter Interpolation." pith.science (2026). https://pith.science/paper/TSVMN3D5

@misc{pith2026250609318,
  author       = {Pith},
  title        = {Pith review of: Canonical Partition Function on a Quantum Computer through Trotter Interpolation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TSVMN3D5}},
  note         = {Machine review of arXiv:2506.09318}
}
abstract

In this work, we present a Gibbs state observable estimation algorithm based on Trotter interpolation, which reaches a state-of-the-art quantum computational cost of $ \tilde{O}(\beta \log{1/\epsilon})$. Our approach saves $\log(\Gamma)$ ancilla qubits compared with the qubitization-based methods for Hamiltonian with $\Gamma$ stages. To provide a robust assessment of our approach, we benchmark our results against state-of-the-art methodology using the SYK model as a testbed. Our method provides an efficient alternative method for Gibbs-state accessing based on Trotterization in the context of quantum state preparation and estimation of thermal observables.

Figures

Figures reproduced from arXiv: 2506.09318 by the authors.

Figure 1
Figure 1. FIG. 1. Quantum circuit for GQSP wherein the angles [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. FIG. 3. ancillary qubits saved compared with the qubitization method [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 2
Figure 2. FIG. 2. GQSP of the function [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 25 canonical work pages

  1. [33]

    Grinko, J

    D. Grinko, J. Gacon, C. Zoufal, and S. Woerner, npj Quantum Information7, 52 (2021)

    work page 2021

  2. [1]

    Duane, A

    S. Duane, A. D. Kennedy, B. J. Pendleton, and D. Roweth, Phys. Lett. B195, 216 (1987)

    1987

  3. [2]

    The theory of hybrid stochastic algorithms,

    A. D. Kennedy, “The theory of hybrid stochastic algorithms,” in Probabilistic Methods in Quantum Field Theory and Quantum Gravity, edited by P. H. Damgaard, H. H¨ uffel, and A. Rosenblum (Springer US, Boston, MA, 1990) pp. 209–223

    work page 1990

  4. [3]

    Piccioli, E

    G. Piccioli, E. Troiani, and L. Zdeborov ´a, Journal of Physics A: Mathematical and Theoretical57, 125002 (2024)

    work page 2024

  5. [4]

    Tomography and Generative Data Modeling via Quantum Boltzmann Training

    M. Kieferov ´a and N. Wiebe, Phys. Rev. A96, 062327 (2017), arXiv:1612.05204 [quant-ph]

    work page Pith review arXiv 2017

  6. [5]

    Biamonte, P

    J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, and S. Lloyd, Nature (London)549, 195 (2017), arXiv:1611.09347 [quant-ph]

    arXiv 2017

  7. [6]

    R. D. Somma, S. Boixo, H. Barnum, and E. Knill, Phys. Rev. Lett.101, 130504 (2008)

    work page 2008

  8. [7]

    F. G. S. L. Brandao and K. Svore, arXiv e-prints , arXiv:1609.05537 (2016), arXiv:1609.05537 [quant-ph]

    arXiv 2016

Show all 43 references
  1. [8]

    van Apeldoorn, A

    J. van Apeldoorn, A. Gily ´en, S. Gribling, and R. de Wolf, Quantum4, 230 (2020)

    2020

  2. [9]

    Poulin and P

    D. Poulin and P. Wocjan, Phys. Rev. Lett.103, 220502 (2009), arXiv:0905.2199 [quant-ph]

    2009 arXiv

  3. [10]

    Kitaev, A simple model of quantum holography, KITP Pro- gram: Entanglement in Strongly-Correlated Quantum Matter (1993)

    A. Kitaev, A simple model of quantum holography, KITP Pro- gram: Entanglement in Strongly-Correlated Quantum Matter (1993)

    1993

  4. [11]

    Sachdev and J

    S. Sachdev and J. Ye, Phys. Rev. Lett.70, 3339 (1993)

    1993

  5. [12]

    A. A. Patel, H. Guo, I. Esterlis, and S. Sachdev, Science381, 790–793 (2023)

    2023

  6. [13]

    Sachdev, Physical Review Letters105(2010), 10.1103/phys- revlett.105.151602

    S. Sachdev, Physical Review Letters105(2010), 10.1103/phys- revlett.105.151602

    2010 doi

  7. [14]

    Chowdhury, A

    D. Chowdhury, A. Georges, O. Parcollet, and S. Sachdev, Re- views of Modern Physics94, 035004 (2022)

    2022

  8. [15]

    Maldacena and D

    J. Maldacena and D. Stanford, Physical Review D94(2016), 10.1103/physrevd.94.106002

    2016 doi

  9. [16]

    J. S. Cotler, G. Gur-Ari, M. Hanada, J. Polchinski, P. Saad, S. H. Shenker, D. Stanford, A. Streicher, and M. Tezuka, Journal of High Energy Physics2017(2017), 10.1007/jhep05(2017)118

    2017 doi

  10. [17]

    Quantum chaos, thermodynamics and black hole microstates in the mass deformed syk model,

    T. Nosaka and T. Numasawa, “Quantum chaos, thermodynamics and black hole microstates in the mass deformed syk model,” (2020), arXiv:1912.12302 [hep-th]

    2020 arXiv

  11. [18]

    Quantum algorithms for gibbs sampling and hitting-time estimation,

    A. N. Chowdhury and R. D. Somma, “Quantum algorithms for gibbs sampling and hitting-time estimation,” (2016), arXiv:1603.02940 [quant-ph]

    2016 arXiv

  12. [19]

    C.-F. Chen, M. J. Kastoryano, F. G. Brand ˜ao, and A. Gily ´en, arXiv preprint arXiv:2303.18224 (2023)

    2023 arXiv

  13. [20]

    An efficient and exact noncommutative quantum gibbs sampler,

    C.-F. Chen, M. J. Kastoryano, and A. Gily ´en, “An efficient and exact noncommutative quantum gibbs sampler,” (2023), arXiv:2311.09207 [quant-ph]

    2023

  14. [21]

    Nelson, M

    J. Nelson, M. Vuffray, A. Y. Lokhov, T. Albash, and C. Coffrin, Phys. Rev. Appl.17, 044046 (2022)

    2022

  15. [22]

    Bergamaschi, C.-F

    T. Bergamaschi, C.-F. Chen, and Y. Liu, in2024 IEEE 65th An- nual Symposium on Foundations of Computer Science (FOCS) (IEEE, 2024) p. 1063–1085

    2024

  16. [23]

    Gibbs sampling gives quantum advantage at constant temperatures with o(1)-local hamiltoni- ans,

    J. Rajakumar and J. D. Watson, “Gibbs sampling gives quantum advantage at constant temperatures with o(1)-local hamiltoni- ans,” (2024), arXiv:2408.01516 [quant-ph]

    2024

  17. [24]

    N. M. Eassa, M. M. Moustafa, A. Banerjee, and J. Cohn, npj Quantum Information10(2024), 10.1038/s41534-024-00887- w

    2024 doi

  18. [25]

    M. F. Sykes, J. W. Essam, B. R. Heap, and B. J. Hiley, Journal of Mathematical Physics7, 1557 (1966), https://pubs.aip.org/aip/jmp/article- pdf/7/9/1557/19097485/1557 1 online.pdf

    1966

  19. [26]

    Oitmaa, C

    J. Oitmaa, C. Hamer, and W. Zheng,Series Expansion Methods for Strongly Interacting Lattice Models(Cambridge University Press, 2006)

    2006

  20. [27]

    Gily ´en, Y

    A. Gily ´en, Y. Su, G. H. Low, and N. Wiebe, inProceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing, STOC ’19 (ACM, 2019) p. 193–204

    2019

  21. [28]

    Quantum many- body simulation of finite-temperature systems with sampling a series expansion of a quantum imaginary-time evolution,

    N. Matsumoto, S. Tsutsui, Y. O. Nakagawa, Y. Hidaka, S. Kana- sugi, K. Maruyama, H. Oshima, and S. Sato, “Quantum many- body simulation of finite-temperature systems with sampling a series expansion of a quantum imaginary-time evolution,” (2024), arXiv:2409.07070 [quant-ph]

    2024 arXiv

  22. [29]

    Quantum Signal Processing method to use to implement the LWF for the function𝑒 −𝛽(𝑥+1) which is needed to imple- menting the Boltzmann factor𝑒−𝛽(𝐻+1) . IV. QUANTUM SIGNAL PROCESSING FOR TROTTERIZED EVOLUTION OPERATOR One method of implementing the Trotterized evolution op- era...

  23. [30]

    Suzuki, Journal of Mathematical Physics32, 400 (1991), https://pubs.aip.org/aip/jmp/article- pdf/32/2/400/19166143/400 1 online.pdf

    M. Suzuki, Journal of Mathematical Physics32, 400 (1991), https://pubs.aip.org/aip/jmp/article- pdf/32/2/400/19166143/400 1 online.pdf

    1991

  24. [31]

    Generalized quantum signal pro- cessing,

    D. Motlagh and N. Wiebe, “Generalized quantum signal pro- cessing,” (2024), arXiv:2308.01501 [quant-ph]

    2024 arXiv

  25. [32]

    Rendon, J

    G. Rendon, J. Watkins, and N. Wiebe, Quantum8, 1266 (2024)

    2024

  26. [34]

    With this, the interpolation is well conditioned as the condition number forVis optimal, i.e

    with respect to the collocation nodes, that is, 𝑀cheb∑︁ 𝑘=1 𝑢𝑖(𝑠𝑘)𝑢𝑗(𝑠𝑘)=𝛿 𝑖𝑗 (12) for all 0≤𝑖,𝑗 < 𝑀 cheb. With this, the interpolation is well conditioned as the condition number forVis optimal, i.e. 𝜅(V)=𝜎 max(V)/𝜎min(V)=1. After simplifications, we can bypass the estimation...

  27. [35]

    Towards dequantizing quantum signal processing,

    G. Rendon, “Towards dequantizing quantum signal processing,” (2023), arXiv:2311.01533 [quant-ph]

    2023 arXiv

  28. [36]

    Mason and D

    J. Mason and D. Handscomb,Chebyshev Polynomials(CRC Press, 2002)

    2002

  29. [37]

    L. N. Trefethen,Approximation Theory and Approximation Practice, Extended Edition(SIAM-Society for Industrial and Applied Mathematics, Philadelphia, PA, USA, 2019)

    2019

  30. [38]

    Quantum amplitude amplification and estimation,

    G. Brassard, P. Høyer, M. Mosca, and A. Tapp, “Quantum amplitude amplification and estimation,” (2002)

    2002

  31. [39]

    Luo, Y.-Z

    Z. Luo, Y.-Z. You, J. Li, C.-M. Jian, D. Lu, C. Xu, B. Zeng, and R. Laflamme, npj Quantum Information5(2019), 10.1038/s41534-019-0166-7

    2019 doi

  32. [40]

    Garc ´ıa- ´Alvarez, I

    L. Garc ´ıa- ´Alvarez, I. Egusquiza, L. Lamata, A. del Campo, J. Sonner, and E. Solano, Physical Review Letters119(2017), 10.1103/physrevlett.119.040501

    2017 doi

  33. [41]

    Asaduzzaman, R

    M. Asaduzzaman, R. G. Jha, and B. Sambasivam, Physical Review D109(2024), 10.1103/physrevd.109.105002

    2024 doi

  34. [42]

    Babbush, D

    R. Babbush, D. W. Berry, and H. Neven, Physical Review A99 (2019), 10.1103/physreva.99.040301

    2019 doi

  35. [43]

    G. H. Low and I. L. Chuang, Quantum3, 163 (2019)

    2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.