Pith. sign in

REVIEW 7 minor 36 references

Classical post-processing cuts quantum ODE solver depth from 1/ε to polylog(1/ε)

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 ↗

Classical step-size extrapolation reduces the maximum single-run circuit depth of a one-ancilla quantum ODE solver from O(1/ε) to O(polylog(1/ε)) without adding ancillae.

T0 review reviewed 2026-07-09 challenge →

load-bearing objection Step-size extrapolation reduces one-ancilla quantum ODE solver depth from O(1/ε) to O(polylog(1/ε)) via a holomorphic adjoint lift that resolves the anti-holomorphicity problem.

arxiv 2607.07389 v1 pith:O2LJMBYW submitted 2026-07-08 quant-ph cs.NAmath.NA

Circuit Depth Reduction of One-Ancilla Quantum Differential Equation Solver via Extrapolation

Di Fang , Justin Park This is my paper
classification quant-ph cs.NAmath.NA PACS 03.67.Ac03.67.Lx
keywords circuitdepthquantumsolverdifferentialepsilonequationone-ancilla
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 reading

This paper addresses a fundamental tension in quantum algorithms for differential equations: circuits that are simple enough for near-term hardware tend to scale poorly in accuracy. A recently introduced one-ancilla quantum ODE solver preserves locality and uses only a single ancilla qubit, but its maximum single-run circuit depth grows as O(1/ε) in the target accuracy ε—a direct consequence of its first-order discretization. The authors show that by running this same circuit at a logarithmic number of coarse, distinct step sizes and then classically extrapolating to the zero-step-size limit (using either Richardson extrapolation or Chebyshev interpolation), the maximum single-run depth drops to O(polylog(1/ε)) without adding ancillae or sacrificing locality. The key technical obstacle is that, unlike Hamiltonian or Lindbladian simulation, the postselected ODE setting produces a non-unitary evolution whose observable signal involves a quadratic pairing of the forward map with its Hilbert-space adjoint. When the step size is analytically continued to the complex plane, the adjoint becomes anti-holomorphic, breaking the tools used in prior extrapolation analyses. The authors resolve this by constructing a holomorphic adjoint lift—an operator that agrees with the usual adjoint on real step sizes but remains holomorphic in the complexified step-size variable. This device yields sharp derivative bounds on the observable as a function of step size, which in turn control the extrapolation bias. The paper also introduces a zero-on-rejection estimator that records failed postselection attempts as zero-valued samples, making the success probability automatically absorbed into the estimator without separate estimation. A matching lower bound shows the resulting sample complexity is optimal for the independent-shot model used.

Core claim

The central discovery is that the observable signal g_O(s) = ⟨u_s(T)|O|u_s(T)⟩, produced by the one-ancilla solver at step size s, is sufficiently smooth as a function of s near zero to permit polynomial extrapolation to the exact value g_O(0), even though the underlying dynamics are non-unitary and postselected. The smoothness is established through a holomorphic adjoint lift: one defines A_eff(z) = z⁻¹ log K_z (the modified generator, holomorphic in complex step size z), and then constructs A_♯_eff(z) by conjugating only the operator coefficients of its Taylor series while leaving z holomorphic. This lift agrees with the true adjoint on real step sizes and remains holomorphic off the real轴

What carries the argument

Holomorphic adjoint lift; zero-on-rejection estimator; Richardson and Chebyshev extrapolation on step-size-dependent observable signals

Load-bearing premise

The derivative bounds require step sizes to lie within a disk of radius proportional to 1/Q (where Q = max{S, T·Γ_loc} combines generator strength and accumulated dissipation), meaning the extrapolation nodes must be small enough for the holomorphic lift to be bounded. If the generator is very strong or the evolution time is very long, the admissible step sizes become very small, which could narrow or erase the depth advantage.

What would settle it

A problem instance where Q = max{S, T·Γ_loc} is so large that the admissible step sizes s ≤ c/Q force the coarsest extrapolation node to require depth comparable to or exceeding the direct O(1/ε) method, erasing the polylogarithmic advantage. Concretely, if T·Γ_loc ≫ 1/ε, the constraint s_max = c/Q may force R₁ = T/s_max so large that the maximum depth d_step·R_m already exceeds what direct estimation at accuracy ε would require.

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

If this is right

  • Maximum single-run circuit depth for the one-ancilla ODE solver improves from O(1/ε) to O(log(1/ε)) (Richardson) or O(log²(1/ε)) (Chebyshev), making high-accuracy observable estimation feasible on early fault-tolerant hardware with shallow circuits.
  • The holomorphic adjoint lift technique may extend to other postselected or non-unitary quantum algorithms where observable estimates involve paired forward-adjoint maps, broadening the class of quantum simulations amenable to step-size extrapolation.
  • The zero-on-rejection estimator design—recording postselection failures as zeros rather than discarding them—decouples sample complexity from success probability, which could simplify resource analysis for other postselected quantum protocols.
  • The trade-off between Richardson (smaller depth, exponentially growing weights) and Chebyshev (larger depth, polylogarithmic weights) provides a tunable design parameter: practitioners can select the method based on whether circuit depth or sampling budget is the binding constraint.
Share X Bluesky LinkedIn Reddit HN

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

0 major / 7 minor

Summary. This paper proposes a classical extrapolation framework to reduce the maximum single-run circuit depth of the one-ancilla quantum ODE solver from O(1/ε) to O(polylog(1/ε)) in target accuracy ε. The approach runs the existing one-ancilla circuit at a logarithmic number of finite step sizes and combines the results via Richardson extrapolation or Chebyshev interpolation. The key technical challenge is obtaining derivative bounds for the observable map g_O(s) under nonunitary evolution, where the Hilbert-space adjoint is anti-holomorphic when the step size is complexified. The authors resolve this by introducing a holomorphic adjoint lift A^#_eff(z) that agrees with the usual adjoint on real step sizes but remains holomorphic for complex z. The proof chain proceeds through local error bounds (Props. 4–5), complex-step extensions (Lemmas 6–9), the holomorphic adjoint lift (Lemma 10), sharp derivative bounds via Cauchy's formula (Theorem 11), deterministic bias bounds (Theorem 12), sampling concentration (Theorem 14), and explicit resource estimates for both extrapolation methods (Theorems 18, 20). A matching lower bound for the independent-shot sampling model is provided (Theorem 28). Numerical experiments on the Hatano-Nelson model and a convection-diffusion equation illustrate the theory.

Significance. The paper makes a genuine contribution by extending step-size extrapolation techniques—previously established for Hamiltonian and Lindbladian simulation—to the postselected, nonunitary setting of the one-ancilla ODE solver. The holomorphic adjoint lift (Section 3.3) is a non-trivial technical innovation that correctly resolves the anti-holomorphicity of the naive adjoint. The extrapolation framework is parameter-free in the relevant sense: step sizes s_ℓ are determined by the target accuracy ε and problem constants Q, S, Γ_loc, and the extrapolation weights a_ℓ are computed from fixed Lagrange interpolation nodes, not fitted post-hoc. The Q-dependence cancellation in the bias bound—where (C'Q)^m from the derivative bound cancels with (c/Q)^m from the step-size product—is the mechanism that makes the depth reduction work and is correctly derived. The matching sampling lower bound (Theorem 28) establishes optimality of the ε^{-2} factor in the independent-shot model. The numerical experiments, while small-scale, are consistent with the theoretical predictions.

minor comments (7)
  1. Section 3.3: The notation A^#_eff(z) may be confused with musical isomorphism notation from differential geometry. A brief remark explaining that '#' here denotes the holomorphic adjoint lift, not a sharp operator, would improve readability.
  2. Theorem 11, Step 1: The regime condition s ≤ s_Q := c/Q with c ≤ min{c_3/4, 1/(2C_3)} is load-bearing for the derivative bound. While the paper notes this restricts admissible step sizes, it would help to state more prominently—perhaps in a dedicated 'Regime of Validity' remark—how this condition interacts with the resource estimates in Theorems 18 and 20, especially for practitioners who may wonder when Q is large.
  3. Section 4.4, Remark 2: The resolution condition T ≥ b_3 s_max m^2 log m for Chebyshev translates to T ≳ (1/√n) log(1/ε) √(log log(1/ε)). This is a mild condition, but it would be useful to briefly discuss whether it is automatically satisfied in typical use cases or whether it imposes a practical constraint.
  4. Section 5: The numerical experiments use very small systems (N=5 for Hatano-Nelson, 6×6 grid for convection-diffusion). While the authors note these are for demonstration purposes, a brief discussion of whether finite-size effects might obscure the asymptotic scaling predictions would strengthen the presentation.
  5. Theorem 18 proof: The assumption s_max ≤ T (equivalently c ≤ QT) is stated to hold when 'T is not too small relative to S or to the dissipation rate Γ_loc.' It would be helpful to state this condition explicitly in the theorem statement rather than only in the proof, since it is a hypothesis for the resource bound.
  6. Figure 1(b) and Figure 3(b): Plotting bias versus √R_max is a sensible way to visualize the Chebyshev scaling, but the caption could explicitly state why this rescaling is chosen (i.e., R_max = Θ(m^2) for Chebyshev).
  7. Reference [21] is cited frequently as the source of the one-ancilla algorithm. Since this appears to be a concurrent or very recent preprint, the authors should verify that the cited results (particularly the algorithm construction and its properties) are accurately represented.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for a careful and positive assessment of our manuscript. The referee's recommendation is minor revision, and the report does not identify any major comments requiring substantive changes. We address the referee's points below.

read point-by-point responses

  1. Referee: The referee report contains no major comments. The referee's summary accurately describes the paper's contributions: the extrapolation framework, the holomorphic adjoint lift, the Q-dependence cancellation mechanism, the matching sampling lower bound, and the numerical experiments.

    Authors: We thank the referee for the thorough and accurate summary of our work. The referee has correctly identified the key technical contributions, including the holomorphic adjoint lift (Section 3.3) as the central innovation resolving the anti-holomorphicity problem, the Q-dependence cancellation in the bias bound, and the role of the matching lower bound (Theorem 28) in establishing optimality of the epsilon^{-2} sampling factor. We are pleased that the referee finds these contributions to be genuine and non-trivial. Since the referee has not raised any specific concerns requiring revision, we have no changes to make at this time. If the referee or editor has any specific suggestions for clarification or additional discussion that would strengthen the manuscript, we would be happy to incorporate them. revision: no

Circularity Check

0 steps flagged

No circularity found

full rationale

The paper's central depth-reduction claim follows from a genuine mathematical cancellation: the derivative bound (Theorem 11) contributes a factor Q^m, while the step-size product in the Lagrange remainder (Theorem 12) contributes (c/Q)^m, and these cancel to give a Q-independent bias bound C||O||(C'c)^m/m!. This cancellation is algebraic, not definitional. The step sizes s_ℓ are determined by the target accuracy ε and problem constants (Q, S, Γ_loc), not fitted to data. The extrapolation weights a_ℓ are computed from Lagrange interpolation at fixed nodes, not optimized post-hoc. Self-citations are present but not load-bearing for the central claim: reference [21] (co-authored by Fang) provides the base one-ancilla algorithm, but the paper re-derives all necessary properties from scratch (Lemmas 1–5, Propositions 4–5). References [25, 26] provide Chebyshev weight stability bounds (Lemma 24) that affect only sample complexity, not the maximum-depth bound. The holomorphic adjoint lift (Section 3.3) is a genuine construction that resolves the anti-holomorphicity problem without circularity. The numerical experiments (Section 5) validate the theory on independent test problems (Hatano-Nelson, convection-diffusion) without fitting parameters to the predicted scaling. The derivation chain is self-contained against external benchmarks, so the circularity score is 0.

Axiom & Free-Parameter Ledger

3 free parameters · 5 axioms · 0 invented entities

The paper introduces no new physical entities, particles, forces, or postulated objects. The holomorphic adjoint lift A♯_eff(z) is a mathematical construction, not a new physical entity. All axioms are standard domain assumptions for dissipative quantum dynamics. The free parameters (s_max, m, N) are determined by the problem constants and target accuracy, not fitted to data.

free parameters (3)
  • s_max = c/Q where Q = max{S, T·Γ_loc}
    Upper bound on the largest extrapolation step size, chosen as c/Q for a universal constant c. Not fitted to data; determined by problem constants.
  • m (number of extrapolation nodes) = O(log(‖O‖/ε))
    Number of step sizes used in extrapolation, chosen to make the deterministic bias ≤ ε/2. Determined by the accuracy target, not fitted.
  • N (samples per node) = ⌈8‖O‖²Λ_m²/ε² · log(2m/δ)⌉
    Number of circuit executions per step size, from Hoeffding's inequality. Determined by accuracy and failure probability targets.
axioms (5)
  • domain assumption The ODE generator A = -iH - Σ D_j with H Hermitian and D_j = L_j† L_j ⪰ 0 (dissipative form)
    Stated in Eq. (1), Section 2.1. This is the standard dissipative setting shared across state-of-the-art quantum ODE frameworks, as noted in the introduction.
  • domain assumption The one-ancilla algorithm of [21] correctly implements the postselected block C_j(s) = cos(√(2s) D_j^{1/2})
    Lemma 1, Section 3.1. The closed form is proved from the Hermitian dilation G_j, but the circuit construction itself is from [21].
  • domain assumption H and each L_j are geometrically local on n qubits (for Corollary 22)
    Section 4.5. Each acts on O(1) qubits on a constant-dimensional lattice, J = O(n), local coefficients O(1). Standard for spatially local quantum systems.
  • domain assumption Independent-shot sampling model: each circuit execution produces one classical measurement outcome
    Section 4.1, Theorem 28. No coherent amplitude estimation or amplitude amplification is used. This is a modeling choice that keeps the circuit near-term friendly.
  • domain assumption Chebyshev weight bound under node rounding (Lemma 24, adapted from [25, Lemma 9])
    Section 4.4, Lemma 24. The condition T ≥ b₃·s_max·m²·log(m) ensures rounded Chebyshev nodes have well-conditioned weights. Imported from prior work on Lindblad extrapolation.

reviewed 2026-07-09 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Circuit Depth Reduction of One-Ancilla Quantum Differential Equation Solver via Extrapolation." pith.science (2026). https://pith.science/paper/O2LJMBYW

@misc{pith2026260707389,
  author       = {Pith},
  title        = {Pith review of: Circuit Depth Reduction of One-Ancilla Quantum Differential Equation Solver via Extrapolation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O2LJMBYW}},
  note         = {Machine review of arXiv:2607.07389}
}
Share X Bluesky LinkedIn Reddit HN
abstract

Solving linear differential equations is a fundamental task in scientific computing and an important primitive for quantum computing. A recent one-ancilla quantum differential equation solver provides a hardware-friendly and locality-preserving approach with provable performance guarantees, making it highly suitable for the early fault-tolerant and near-term regimes. Its simple circuit structure comes with a natural trade-off: the maximum single-run circuit depth scales as $O (1/\epsilon)$ in the target accuracy $\epsilon$. In this work, we reduce this depth by combining the solver with classical step-size postprocessing. By running the one-ancilla solver at a logarithmic number of finite time step sizes and using classical post-processing to cancel leading discretization errors, we reduce the maximum single-run circuit depth to $O(\mathrm{polylog}(1/\epsilon))$ without adding quantum ancillae or sacrificing locality. Technically, extending extrapolation ideas beyond Hamiltonian and Lindbladian dynamics requires regularity estimates for observable maps under nonunitary evolution, which we obtain through a holomorphic extension of the adjoint evolution. Numerical experiments on the Hatano-Nelson model (ODE) and the convection-diffusion equation (PDE) demonstrate the effectiveness of the approach.

Figures

Figures reproduced from arXiv: 2607.07389 by Di Fang, Justin Park.

Figure 1
Figure 1. Figure 1: Hatano-Nelson model with edge observable. Both panels are semilogy plots: the y-axis is the absolute deterministic bias on a logarithmic scale, while the x-axis is linear. Panel (a) plots the data against 𝑅max, and panel (b) plots the same data against √ 𝑅max. The direct method without the proposed postprocessing requires significantly larger circuit depth, while Richardson extrapolation and Chebyshev redu… view at source ↗
Figure 2
Figure 2. Figure 2: Hatano-Nelson edge observable. This is a log-log plot: the x-axis is the target inverse accuracy 1∕𝜖, and the y-axis is the smallest observed maximum step number 𝑅max needed to reach the target deterministic bias. The direct method follows the expected inverse-accuracy trend, while the extrapolated methods require much smaller maximum depths over the tested accuracy range. The final upward step in the Cheb… view at source ↗
Figure 3
Figure 3. Figure 3: Convection-diffusion low-Fourier-shell observable. Both panels are semilogy plots: the y-axis is the absolute deterministic bias on a logarithmic scale, while the x-axis is linear. Panel (a) plots the data against 𝑅max, and panel (b) plots the same data against √ 𝑅max. For Richardson and Chebyshev, the plotted degrees are 𝑚 = 2,… , 9. The extrapolated curves show rapid deterministic bias reduction relative… view at source ↗
Figure 4
Figure 4. Figure 4: Convection-diffusion low-Fourier-shell observable. This is a log-log plot: the x-axis is the target inverse accuracy 1∕𝜖, and the y-axis is the smallest observed maximum step number 𝑅max needed to reach the target deterministic bias. The direct method requires rapidly increasing 𝑅max as the target accuracy is tightened, while Richardson extrapolation and Chebyshev interpolation remain much lower over the t… 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

36 extracted references · 36 canonical work pages · 7 internal anchors

  1. [1]

    High-order quantum algorithm for solving linear differential equations.Journal of Physics A: Mathematical and Theoretical, 47(10):105301, feb 2014

    Dominic W Berry. High-order quantum algorithm for solving linear differential equations.Journal of Physics A: Mathematical and Theoretical, 47(10):105301, feb 2014

    work page 2014

  2. [2]

    Quantum algorithm for linear differential equations with exponentially improved dependence on precision

    Dominic W. Berry, Andrew M. Childs, Aaron Ostrander, and Guoming Wang. Quantum algorithm for linear differential equations with exponentially improved dependence on precision.Communications in Mathematical Physics, 356(3):1057–1081, 2017. arXiv:1701.03684

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  3. [3]

    QuantumSpectralMethodsforDifferentialEquations.Commun.Math.Phys.,375(2):1427–1457,2020

    AndrewM.ChildsandJinPengLiu. QuantumSpectralMethodsforDifferentialEquations.Commun.Math.Phys.,375(2):1427–1457,2020

    work page 2020

  4. [4]

    Improved quantum algorithms for linear and nonlinear differential equations.Quantum, 7:913, February 2023

    Hari Krovi. Improved quantum algorithms for linear and nonlinear differential equations.Quantum, 7:913, February 2023

    work page 2023

  5. [5]

    Quantum algorithm for time-dependent differential equations using dyson series.Quantum, 8:1369, 2024

    Dominic W Berry and Pedro CS Costa. Quantum algorithm for time-dependent differential equations using dyson series.Quantum, 8:1369, 2024

    work page 2024

  6. [6]

    Fast-forwarding quantum algorithms for linear dissipative differential equations.Quantum, 10:1986, January 2026

    Dong An, Akwum Onwunta, and Gengzhi Yang. Fast-forwarding quantum algorithms for linear dissipative differential equations.Quantum, 10:1986, January 2026

    work page 1986

  7. [7]

    Time-marching based quantum solvers for time-dependent linear differential equations.Quantum, 7:955, 2023

    Di Fang, Lin Lin, and Yu Tong. Time-marching based quantum solvers for time-dependent linear differential equations.Quantum, 7:955, 2023

    work page 2023

  8. [8]

    Quantum Differential Equation Solvers with Low State Preparation Cost: Eliminating the Time Dependence in Dissipative Equations

    Gengzhi Yang, Akwum Onwunta, and Dong An. Quantum differential equation solvers with low state preparation cost: Eliminating the time dependence in dissipative equations, 2025. arXiv:2508.15170

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  9. [9]

    Linear combination of hamiltonian simulation for nonunitary dynamics with optimal state preparation cost.Physical Review Letters, 131(15):150603, 2023

    Dong An, Jin-Peng Liu, and Lin Lin. Linear combination of hamiltonian simulation for nonunitary dynamics with optimal state preparation cost.Physical Review Letters, 131(15):150603, 2023

    work page 2023

  10. [10]

    Laplace transform based quantum eigenvalue transformation via linear combination of Hamiltonian simulation

    Dong An, Andrew M Childs, Lin Lin, and Lexing Ying. Laplace transform based quantum eigenvalue transformation via linear combination of hamiltonian simulation.arXiv preprint arXiv:2411.04010, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  11. [11]

    arXiv preprint arXiv:2312.03916, 2023

    DongAn,AndrewMChilds,andLinLin.Quantumalgorithmforlinearnon-unitarydynamicswithnear-optimaldependenceonallparameters. arXiv preprint arXiv:2312.03916, 2023

    work page arXiv 2023

  12. [12]

    Novikau and I

    I. Novikau and I. Joseph. Quantum algorithm for the advection-diffusion equation and the koopman-von neumann approach to nonlinear dynamical systems.Computer Physics Communications, 309:109498, April 2025

    work page 2025

  13. [13]

    Johnson, Amara Katabarwa, and Nathan Wiebe

    Matthew Pocrnic, Peter D. Johnson, Amara Katabarwa, and Nathan Wiebe. Constant-factor improvements in quantum algorithms for linear differential equations, 2025. arXiv:2506.20760

    work page arXiv 2025

  14. [14]

    Provable Quantum Speedups for Reaction-Rate Estimation in High-Dimensional Fokker-Planck Dynamics

    TylerKharazi,AhmadM.Alkadri,KranthiK.Mandadapu,andK.BirgittaWhaley. ProvableQuantumSpeedupsforReaction-RateEstimation in High-Dimensional Fokker-Planck Dynamics, 2026. arXiv:2601.15523

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  15. [15]

    From linear differential equations to unitaries: A moment-matching dilation framework with near-optimal quantum algorithms

    Xiantao Li. From linear differential equations to unitaries: A moment-matching dilation framework with near-optimal quantum algorithms. PRX Quantum, 7:020350, Jun 2026

    work page 2026

  16. [16]

    Universal dilation of linear itô sdes: Quantum trajectories and lindblad simulation of second moments,

    Hsuan-Cheng Wu and Xiantao Li. Universal dilation of linear itô sdes: Quantum trajectories and lindblad simulation of second moments,

    work page

  17. [17]

    Quantumsimulationofpartialdifferentialequationsviaschrödingerization.PhysicalReviewLetters,133(23), December 2024

    ShiJin,NanaLiu,andYueYu. Quantumsimulationofpartialdifferentialequationsviaschrödingerization.PhysicalReviewLetters,133(23), December 2024

    work page 2024

  18. [18]

    On the schrödingerization method for linear non-unitary dynamics with optimal dependence on matrix queries, 2025

    Shi Jin, Nana Liu, Chuwen Ma, Yizhe Peng, and Yue Yu. On the schrödingerization method for linear non-unitary dynamics with optimal dependence on matrix queries, 2025. arXiv:2505.00370

    work page arXiv 2025

  19. [19]

    Design nearly optimal quantum algorithm for linear differential equations via lindbladians, 2024

    Zhong-Xia Shang, Naixu Guo, Dong An, and Qi Zhao. Design nearly optimal quantum algorithm for linear differential equations via lindbladians, 2024. arXiv:2410.19628

    work page arXiv 2024

  20. [20]

    Quantumeigenvalueprocessing

    GuangHaoLowandYuanSu. Quantumeigenvalueprocessing. In2024IEEE65thAnnualSymposiumonFoundationsofComputerScience (FOCS), page 1051–1062. IEEE, October 2024

    work page 2024

  21. [21]

    Qubit-Efficient Quantum Algorithm for Linear Differential Equations

    Di Fang, David Lloyd George, and Yu Tong. Qubit-efficient quantum algorithm for linear differential equations, 2025. arXiv:2507.16995

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  22. [22]

    Improved accuracy for trotter simulations using chebyshev interpolation.Quantum, 8:1266, February 2024

    Gumaro Rendon, Jacob Watkins, and Nathan Wiebe. Improved accuracy for trotter simulations using chebyshev interpolation.Quantum, 8:1266, February 2024

    work page 2024

  23. [23]

    Exponentially Reduced Circuit Depths Using Trotter Error Mitigation

    James D. Watson and Jacob Watkins. Exponentially reduced circuit depths using trotter error mitigation, 2024. arXiv:2408.14385

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  24. [24]

    James D. Watson. Randomly compiled quantum simulation with exponentially reduced circuit depths, 2024. arXiv:2411.04240

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  25. [25]

    DirectAnalysisofZero-NoiseExtrapolation:PolynomialMethods,ErrorBounds,andSimultaneous Physical-Algorithmic Error Mitigation.Quantum, 9:1909, November 2025

    PegahMohammadipourandXiantaoLi. DirectAnalysisofZero-NoiseExtrapolation:PolynomialMethods,ErrorBounds,andSimultaneous Physical-Algorithmic Error Mitigation.Quantum, 9:1909, November 2025

    work page 1909

  26. [26]

    Reducing circuit depth in lindblad simulation via step-size extrapolation, 2026

    Pegah Mohammadipour and Xiantao Li. Reducing circuit depth in lindblad simulation via step-size extrapolation, 2026. arXiv:2507.22341

    work page arXiv 2026

  27. [27]

    Gambetta

    Kristan Temme, Sergey Bravyi, and Jay M. Gambetta. Error mitigation for short-depth quantum circuits.Physical Review Letters, 119(18), November 2017

    work page 2017

  28. [28]

    Benjamin

    Ying Li and Simon C. Benjamin. Efficient variational quantum simulator incorporating active error minimization.Physical Review X, 7(2), 2017

    work page 2017

  29. [29]

    Benjamin, and Ying Li

    Suguru Endo, Simon C. Benjamin, and Ying Li. Practical quantum error mitigation for near-future applications.Physical Review X, 8(3), 2018

    work page 2018

  30. [30]

    Benjamin, Suguru Endo, William J

    Zhenyu Cai, Ryan Babbush, Simon C. Benjamin, Suguru Endo, William J. Huggins, Ying Li, Jarrod R. McClean, and Thomas E. O’Brien. Quantum error mitigation.Reviews of Modern Physics, 95(4), December 2023

    work page 2023

  31. [31]

    Childs, Yuan Su, Minh C

    Andrew M. Childs, Yuan Su, Minh C. Tran, Nathan Wiebe, and Shuchen Zhu. Theory of Trotter Error with Commutator Scaling.Phys. Rev. X, 11(1):11020, 2021

    work page 2021

  32. [32]

    Time-dependentunboundedHamiltoniansimulationwithvectornormscaling.Quantum,5:1–49,may2021

    DongAn,DiFang,andLinLin. Time-dependentunboundedHamiltoniansimulationwithvectornormscaling.Quantum,5:1–49,may2021

    work page

  33. [33]

    Nonparametric estimators

    Alexandre B Tsybakov. Nonparametric estimators. InIntroduction to Nonparametric Estimation, pages 1–76. Springer, 2008

    work page 2008

  34. [34]

    Localizationtransitionsinnon-hermitianquantummechanics.Phys.Rev.Lett.,77:570–573,Jul1996

    NaomichiHatanoandDavidR.Nelson. Localizationtransitionsinnon-hermitianquantummechanics.Phys.Rev.Lett.,77:570–573,Jul1996

    work page

  35. [35]

    Naomichi Hatano and David R. Nelson. Vortex pinning and non-hermitian quantum mechanics.Phys. Rev. B, 56:8651–8673, Oct 1997

    work page 1997

  36. [36]

    Naomichi Hatano and David R. Nelson. Non-hermitian delocalization and eigenfunctions.Phys. Rev. B, 58:8384–8390, Oct 1998. D. Fang and J. Park:Preprint Page 28 of 28

    work page 1998

This paper was first reviewed by glm-5.2 on July 9, 2026.