arxiv: 2604.20513 · v2 · submitted 2026-04-22 · 🧮 math.NA · cs.NA· quant-ph

Recognition: unknown

Constrained Optimal Polynomials for Quantum Linear System Solvers

Matthias Deiml , Daniel Peterseim

Authors on Pith no claims yet

Pith reviewed 2026-05-09 23:51 UTC · model grok-4.3

classification 🧮 math.NA cs.NAquant-ph

keywords quantum linear systemspolynomial approximationQSVTblock encodingmaximum entropynumerical stabilityKrylov methodsnoise robustness

0 comments

The pith

Constrained optimal polynomials let quantum linear system solvers achieve lower error than standard QSVT and Chebyshev methods by tailoring the approximation to spectral models.

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

The paper presents constrained optimal polynomials as a way to realize the inverse map in quantum linear system solvers at lower polynomial degree while respecting the normalization limits of block encodings. Two concrete families are developed: constrained uniform polynomials that optimize the accuracy-normalization tradeoff under a uniform spectral model consistent with known bounds, and constrained adaptive polynomials that replace the uniform model with a maximum-entropy reconstruction of the eigenvalue measure extracted from QSVT measurements. Numerical experiments under both hardware and stochastic noise demonstrate that these constructions produce smaller solution errors than conventional QSVT-based and Chebyshev-iteration solvers, with the advantage most pronounced when noise dominates the total error. The approach therefore targets the practical bottleneck in near-term quantum linear solvers where circuit depth and noise accumulation limit performance.

Core claim

Constrained Uniform Polynomial solvers optimize a chosen error measure subject to the block-encoding normalization constraint under a uniform spectral model, while Constrained Adaptive Polynomial solvers retain the same constrained optimization structure but substitute a maximum-entropy probability measure reconstructed from spectral moments obtained via QSVT. Both families therefore produce polynomial transformations that balance approximation quality against implementation cost more effectively than unconstrained or fixed-basis alternatives.

What carries the argument

Constrained optimal polynomials obtained by minimizing a chosen error functional subject to an upper bound on the polynomial norm that is consistent with the block-encoding normalization factor.

If this is right

CUP solvers deliver robust error reduction for generic spectra without requiring detailed prior information about the eigenvalues.
CAP solvers can deliver further improvement whenever the spectrum possesses exploitable structure that can be captured by the moment reconstruction.
Both methods show their largest gains relative to standard solvers in the noise-limited regime that is typical of current hardware.
The framework inherits classical Krylov-subspace ideas to guide the choice of polynomial under the quantum block-encoding constraints.

Where Pith is reading between the lines

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

Lower polynomial degrees may become feasible for a target accuracy, which would directly reduce the number of block-encoding applications and therefore the total circuit depth.
The moment-reconstruction technique could be reused in other quantum algorithms that already perform QSVT to obtain cheap spectral information without additional measurements.
The constrained-optimization viewpoint might be extended to other polynomial-based quantum primitives such as Hamiltonian simulation or phase estimation.

Load-bearing premise

The uniform spectral model used for CUP is consistent with the actual eigenvalue bounds, and the maximum-entropy ansatz used for CAP accurately recovers the underlying probability measure from the extracted moments.

What would settle it

A direct comparison on a quantum device or noise simulator with known spectrum in which the observed solution error of CUP or CAP remains equal to or larger than that of standard QSVT or Chebyshev solvers across a range of noise strengths would falsify the claimed performance advantage.

Figures

Figures reproduced from arXiv: 2604.20513 by Daniel Peterseim, Matthias Deiml.

**Figure 1.** Figure 1: Optimal polynomials with 𝑑 = 10. The 𝑦 coordinate of the dots corresponds to the eigenvalues, their size to the weight (𝑉 ⊤𝑏)𝑗 . The polynomial for𝑚 = 9 interpolates at the eigenvalues. The polynomial for𝑚 = 2 approximates the extremal eigenvalues closely, as well as the eigenvalue 𝜆2 having large weight. Optimal linear dependence on 𝜅 can be recovered by approaches based on variable-time techniques [6, 7]… view at source ↗

**Figure 2.** Figure 2: Constructing the polynomial of the Chebyshev iteration, the QSVT, Cheb2, and Cheb3 solvers for 𝑛 = 4. The polynomial in (b) is 0 at 𝑦 = 0, and so division by 𝑦 yields a polynomial. At the same time, it is close to 1 for 𝑦 ∈ [𝜆min, 𝜆max], meaning 1/𝑦 is well approximated in that interval. the absence of any further information about 𝐴, see e.g. [18, Sec. 12.3.2]. It corresponds to the interpolation of 1/𝑦 w… view at source ↗

**Figure 3.** Figure 3: Comparison of different approximations to 1/𝑦 with 𝜆min = 0.1 and 𝜆max = 1, as well as polynomial degree 𝑚 = 3. The Chebyshev iteration offers the best approximation, but is not constrained at all for negative values. The Cheb2 solver fixes this issue, but while the approximation for 𝑦 = 1 is good, it is far from optimal for the smallest eigenvalue. The Cheb3 polynomial balances the approximation at 𝜆min a… view at source ↗

**Figure 4.** Figure 4: Comparison of different approximations to 1/𝑦 with 𝜆min = 0.1 and 𝜆max = 1, as well as polynomial degree 𝑚 = 3. The size of the circles indicates the weight (𝑉 ⊤𝑏)𝑗 of the eigenvalues as given in (4). The CG polynomial well approximates 1/𝑦 at the clusters of eigenvalues, but increases quickly for negative values. The CUP solver has no knowledge of the distribution of eigenvalues and aims at uniform approx… view at source ↗

**Figure 5.** Figure 5: Relationship of median error and complexity for the QSVT, Cheb2, Cheb3, CUP, and CAP solvers, with noise parameter 𝜉 = 0 (left), 𝜉 = 0.0025 (middle), and 𝜉 = 0.005 (right), as well as 𝑁 = 16 · 104 samples per measurement. See Definition 2 for the exact definition of complexity and error. Methods are tested on randomly generated equations with condition number 𝜅˜ = 3. further. Cheb2 seems to have slightly b… view at source ↗

**Figure 6.** Figure 6: Comparison of the QSVT, Cheb2, Cheb3, CUP, and CAP solvers with 𝜉 = 0 (left), 𝜉 = 0.0025 (middle), and 𝜉 = 0.005 (right), as well as 𝑁 = 16 · 104 . The reported measure is the error from Definition 2 relative to the norm of the true solution. The dotted line indicates the 95th percentile of observed errors. that suppresses it outside the spectral interval, as suggested in [4]. However, this rectangle funct… view at source ↗

read the original abstract

Quantum linear system solvers typically realize the inverse map as a polynomial transformation of the spectrum, so their practical cost hinges on implementing this transformation at a low polynomial degree. We introduce constrained optimal polynomials as a framework for this task, drawing on classical Krylov subspace theory. Within this framework, we develop two classes of solvers. Constrained Uniform Polynomial (CUP) solvers optimize the tradeoff between approximation accuracy and block encoding normalization under a uniform spectral model consistent with the available bounds. Constrained Adaptive Polynomial (CAP) solvers retain this structure but replace the uniform model with a probability measure reconstructed from spectral moments via a maximum entropy ansatz, where the moments are extracted from QSVT measurements. Numerical experiments under hardware and stochastic noise show that these methods achieve lower error than standard QSVT-based and Chebyshev-iteration-type solvers, particularly in noise-limited regimes. CUP offers robust performance under generic spectra, while CAP provides further improvement when the spectral structure can be exploited.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper gives a new constrained-optimization framework for polynomials in quantum linear solvers, splitting into a uniform-bound version and an adaptive max-entropy version that pulls moments from QSVT.

read the letter

The main thing to know is that the authors adapt classical Krylov-style polynomial selection to the specific constraints of quantum block encodings. They define constrained uniform polynomials that optimize under a flat spectral model consistent with known bounds, and constrained adaptive polynomials that replace the flat model with a maximum-entropy reconstruction built from moments measured by QSVT. Numerical tests under hardware and stochastic noise then show lower error than plain QSVT or Chebyshev approaches, especially when noise dominates. This is genuinely new in how it folds the block-encoding normalization directly into the objective and lets the polynomial adapt to measured spectral information rather than relying on worst-case bounds alone. The framework is cleanly stated and the distinction between the two classes is useful. The experiments are presented as evidence that the adaptive route can exploit structure when it exists. The soft spot is the maximum-entropy step in the adaptive case. If the true eigenvalue distribution contains gaps or sharp clusters that a small number of moments cannot resolve, the reconstructed measure will be wrong and the resulting polynomial will optimize the wrong target; the claimed gains may then shrink or disappear on spectra that differ from the test cases. The uniform version is less exposed to this issue but also less powerful. The abstract gives no details on the test matrices, the number of moments used, or the statistical variability of the noise runs, so the strength of the performance claim is still hard to gauge. The work is aimed at people who implement or analyze polynomial-based quantum linear solvers and want to reduce degree or error on near-term hardware. Readers already working on QSVT or block-encoding refinements will see the most direct value. The ideas are coherent enough and the claims specific enough that the paper deserves a serious referee, though reviewers should be asked to examine the reconstruction accuracy and the experimental controls in detail. I would send it to review.

Referee Report

3 major / 2 minor

Summary. The manuscript introduces constrained optimal polynomials as a framework for quantum linear system solvers, drawing on classical Krylov subspace methods. It develops two solvers: Constrained Uniform Polynomial (CUP) solvers that optimize the accuracy-normalization tradeoff under a uniform spectral model consistent with available bounds, and Constrained Adaptive Polynomial (CAP) solvers that replace the uniform model with a maximum-entropy reconstruction of the spectral probability measure from moments extracted via QSVT measurements. Numerical experiments under hardware and stochastic noise are reported to show lower error than standard QSVT-based and Chebyshev-iteration solvers, with CUP providing robust generic performance and CAP offering further gains when spectral structure is exploitable.

Significance. If the central performance claims are substantiated, the work offers a principled way to incorporate spectral bounds or measured moments into polynomial design for quantum linear solvers, potentially lowering effective degree or residual error in noise-limited regimes. The explicit link to constrained optimization from classical Krylov theory and the use of real QSVT data for adaptation are positive features that could influence practical implementations.

major comments (3)

[§3.2] §3.2 (CAP construction): The maximum-entropy ansatz is used to reconstruct the spectral measure from QSVT-extracted moments and is load-bearing for the headline claim that CAP outperforms standard solvers when structure is present. No error bounds, convergence rates, or sensitivity analysis are provided for how reconstruction error (e.g., for gapped or clustered spectra) propagates into the optimized polynomial and final solver error.
[Numerical experiments] Numerical experiments section: The abstract and experiments assert lower error than QSVT and Chebyshev baselines under hardware/stochastic noise, yet the manuscript supplies neither the full experimental protocol, the precise matrices or spectra tested, error-bar statistics, nor raw data. This prevents independent verification of the performance advantage and leaves open the possibility that reported gains depend on post-hoc choices or unstated assumptions about the noise model.
[§2] §2 (uniform spectral model): The claim that the uniform model is 'consistent with the available bounds' for CUP is central to its robustness guarantee, but the manuscript does not derive or cite an explicit relation between the block-encoding normalization factor, the polynomial degree, and the resulting approximation error under this model.

minor comments (2)

Notation for the block-encoding normalization factor and the constrained optimization objective is introduced without a consolidated table or clear cross-reference, making it difficult to track across CUP and CAP definitions.
The manuscript would benefit from an explicit statement of the precise polynomial optimization problem (objective and constraints) in equation form before the algorithmic descriptions.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive review and positive assessment of the potential impact of the constrained optimal polynomial framework. We address each major comment point by point below, indicating where we will revise the manuscript.

read point-by-point responses

Referee: [§3.2] §3.2 (CAP construction): The maximum-entropy ansatz is used to reconstruct the spectral measure from QSVT-extracted moments and is load-bearing for the headline claim that CAP outperforms standard solvers when structure is present. No error bounds, convergence rates, or sensitivity analysis are provided for how reconstruction error (e.g., for gapped or clustered spectra) propagates into the optimized polynomial and final solver error.

Authors: We acknowledge that the manuscript does not supply a full theoretical error analysis for the maximum-entropy reconstruction step. The ansatz is a classical, well-studied method for recovering a positive measure from moments, and our contribution centers on embedding it inside the constrained polynomial optimization rather than on new bounds for the reconstruction itself. Nevertheless, to address the concern, we will add a paragraph in the revised §3.2 discussing the stability of the maximum-entropy solution under moment perturbations (with reference to the classical literature) and will include a brief numerical sensitivity study showing how reconstruction error affects the final solver residual for representative gapped and clustered spectra. revision: partial
Referee: [Numerical experiments] Numerical experiments section: The abstract and experiments assert lower error than QSVT and Chebyshev baselines under hardware/stochastic noise, yet the manuscript supplies neither the full experimental protocol, the precise matrices or spectra tested, error-bar statistics, nor raw data. This prevents independent verification of the performance advantage and leaves open the possibility that reported gains depend on post-hoc choices or unstated assumptions about the noise model.

Authors: We agree that the current experimental section lacks sufficient detail for independent reproduction. Section 4 outlines the noise models and the general QSVT-based measurement protocol, but we will expand it to list the exact test matrices (dimensions, spectra, and sources), the precise block-encoding normalizations, the number of independent runs used for statistics, and the full set of hyperparameters. Error bars will be added to all reported figures, and we will deposit the raw data, simulation code, and hardware run logs in a public repository with a DOI link in the revised manuscript. This will eliminate any ambiguity regarding post-hoc choices or noise-model assumptions. revision: yes
Referee: [§2] §2 (uniform spectral model): The claim that the uniform model is 'consistent with the available bounds' for CUP is central to its robustness guarantee, but the manuscript does not derive or cite an explicit relation between the block-encoding normalization factor, the polynomial degree, and the resulting approximation error under this model.

Authors: The uniform spectral model is defined in §2 by taking the support of the measure to be the interval [0,1/α] where α is the block-encoding normalization factor; this choice is consistent with the known singular-value bounds by construction. The constrained optimization then minimizes the L2 error with respect to this uniform measure subject to the normalization constraint on the polynomial. We will revise §2 to state this relation explicitly, derive the resulting worst-case error bound in terms of α and degree, and add citations to the classical theory of minimax and least-squares polynomial approximation on bounded intervals. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation uses external measurements and explicit modeling choices

full rationale

The CUP construction optimizes under an explicit uniform spectral model tied to stated bounds, without fitting to the target error metric. CAP extracts moments from actual QSVT measurements (external data) and applies a standard maximum-entropy reconstruction to obtain the measure for optimization; this supplies independent input rather than deriving the result by construction. Numerical experiments compare performance against QSVT and Chebyshev baselines under hardware/stochastic noise, providing falsifiable external validation. No self-definitional reductions, fitted inputs renamed as predictions, load-bearing self-citations, or ansatz smuggling appear in the framework. The claimed gains rest on the constrained optimization and empirical tests rather than tautological equivalence to inputs.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claims rest on two modeling assumptions whose validity is not derived inside the paper: a uniform spectral model for CUP and a maximum-entropy reconstruction from measured moments for CAP. No new physical entities are postulated.

free parameters (2)

polynomial degree
Chosen to balance approximation accuracy against normalization; specific values are selected or optimized per instance.
block-encoding normalization factor
Optimized under the chosen spectral model rather than fixed a priori.

axioms (2)

domain assumption Uniform spectral model consistent with available bounds
Invoked for CUP solvers to define the optimization objective.
domain assumption Maximum-entropy ansatz reconstructs the spectral measure from QSVT moments
Invoked for CAP solvers to obtain an adaptive probability measure.

pith-pipeline@v0.9.0 · 5462 in / 1417 out tokens · 32891 ms · 2026-05-09T23:51:01.827707+00:00 · methodology

discussion (0)

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Unitaria: Quantum Linear Algebra via Block Encodings
quant-ph 2026-05 accept novelty 4.0

Unitaria is a new open-source Python library that provides a high-level, composable interface for block encodings in quantum computing, enabling automatic circuit generation and classical simulation-based verification.

Reference graph

Works this paper leans on

33 extracted references · 25 canonical work pages · cited by 1 Pith paper · 2 internal anchors

[1]

Harrow and Avinatan Hassidim and Seth Lloyd , Date-Added =

A. W. Harrow, A. Hassidim, and S. Lloyd. “Quantum algorithm for solving linear systems of equations”. Phys. Rev. Lett.103, 150502 (2009). arXiv:0811.3171

work page arXiv 2009
[2]

Childs, Robin Kothari, and Rolando D

A. M. Childs, R. Kothari, and R. D. Somma. “Quantum Algorithm for Systems of Linear Equations with Exponentially Improved Dependence on Precision”. SIAM J. Comput.46, 1920–1950 (2017). arXiv:1511.02306

work page arXiv 1920
[3]

The Power of Block-Encoded Matrix Powers: Improved Regression Techniques via Faster Hamiltonian Simulation

S. Chakraborty, A. Gilyén, and S. Jeffery. “The Power of Block-Encoded Matrix Powers: Improved Regression Techniques via Faster Hamiltonian Simulation”. LIPIcs Vol. 132 ICALP 2019132, 33:1–33:14 (2019). arXiv:1804.01973

work page arXiv 2019
[4]

Quantum singular value transformation and beyond: exponential improvements for quantum matrix arithmetics,

A. Gilyén, Y. Su, G. H. Low, and N. Wiebe. “Quantum singular value transformation and beyond: Exponential improvements for quantum matrix arithmetics”. In Proc. 51st Annu. ACM SIGACT Symp. Theory Comput. Pages 193–204. (2019). arXiv:1806.01838

work page arXiv 2019
[5]

Quantum linear system solvers: A survey of algorithms and applications

M. E. S. Morales, L. Pira, P. Schleich, K. Koor, P. C. S. Costa, D. An, A. Aspuru-Guzik, L. Lin, P. Rebentrost, and D. W. Berry. “Quantum Linear System Solvers: A Survey of Algorithms and Applications” (2025). arXiv:2411.02522

work page arXiv 2025
[6]

Variable time amplitude amplification and a faster quan- tum algorithm for solving systems of linear equations,

A. Ambainis. “Variable time amplitude amplification and a faster quantum algorithm for solving systems of linear equations” (2010). arXiv:1010.4458

work page arXiv 2010
[7]

Quantum linear system algorithm with optimal queries to initial state preparation

G. H. Low and Y. Su. “Quantum linear system algorithm with optimal queries to initial state preparation”. Quantum10, 2041 (2026). arXiv:2410.18178

work page arXiv 2041
[8]

Lin and Y

L. Lin and Y. Tong. “Optimal polynomial based quantum eigenstate filtering with application to solving quantum linear systems”. Quantum4, 361 (2020). arXiv:1910.14596

work page arXiv 2020
[9]

Quantum linear system solver based on time-optimal adiabatic quantum computing and quantum approximate optimization algorithm

D. An and L. Lin. “Quantum linear system solver based on time-optimal adiabatic quantum computing and quantum approximate optimization algorithm”. ACM Transactions on Quantum Computing3, 1–28 (2022). arXiv:1909.05500

work page arXiv 2022
[10]

A shortcut to an optimal quantum linear system solver

A. M. Dalzell. “A shortcut to an optimal quantum linear system solver” (2024). arXiv:2406.12086

work page internal anchor Pith review Pith/arXiv arXiv 2024
[11]

Numerical Determination of Fundamental Modes

D. A. Flanders and G. Shortley. “Numerical Determination of Fundamental Modes”. J. Appl. Phys.21, 1326–1332 (1950)

1950
[12]

Krylov Subspace Methods: Principles and Analysis

J. Liesen and Z. Strakos. “Krylov Subspace Methods: Principles and Analysis”. Oxford University Press. (2012)

2012
[13]

An Optimal Linear-combination-of-unitaries-based Quantum Linear System Solver

S. Gribling, I. Kerenidis, and D. Szilágyi. “An Optimal Linear-combination-of-unitaries-based Quantum Linear System Solver”. ACM Trans. Quantum Comput.5, 1–23 (2024)

2024
[14]

Matrix inversion polynomials for the quantum singular value transformation

C. Sünderhauf, Z. Németh, A. Walayat, A. Patterson, and B. K. Berntson. “Matrix inversion polynomials for the quantum singular value transformation” (2025). arXiv:2507.15537

work page arXiv 2025
[15]

Quantum conjugate gradient method using the positive-side quantum eigenvalue transformation

K. Toyoizumi, K. Wada, N. Yamamoto, and K. Hoshino. “Quantum conjugate gradient method using the positive-side quantum eigenvalue transformation” (2024). arXiv:2404.02713

work page arXiv 2024
[16]

Quantum Krylov-Subspace Method Based Linear Solver

R.-B. Xu, Z.-J. Zheng, and Z. Zheng. “Quantum Krylov-Subspace Method Based Linear Solver” (2024). arXiv:2405.06359

work page arXiv 2024
[17]

Generalized quantum singular value transformation with application in quantum conjugate gradient least squares algorithm

Y.-Q. Liu, H. Wang, and H. Xiang. “Generalized quantum singular value transformation with application in quantum bi-conjugate gradient method” (2025). arXiv:2508.21390

work page internal anchor Pith review Pith/arXiv arXiv 2025
[18]

Iterative Methods for Sparse Linear Systems

Y. Saad. “Iterative Methods for Sparse Linear Systems”. Society for Industrial and Applied Mathematics. (2003). Second edition

2003
[19]

Maximum entropy in the problem of moments

L. R. Mead and N. Papanicolaou. “Maximum entropy in the problem of moments”. J. Math. Phys.25, 2404–2417 (1984)

1984
[20]

Nonlinear quantum computing by amplified encodings

M. Deiml and D. Peterseim. “Nonlinear quantum computing by amplified encodings” (2024). arXiv:2411.16435. CONSTRAINED OPTIMAL POLYNOMIALS FOR QUANTUM LINEAR SOLVERS 19

work page arXiv 2024
[21]

Quantum Enhanced Numerical Homogeniza- tion

L. Balazi, M. Deiml, and D. Peterseim. “Quantum Enhanced Numerical Homogeniza- tion” (2026). arXiv:2603.28521

work page arXiv 2026
[22]

A single quantum cannot be cloned

W. K. Wootters and W. H. Zurek. “A single quantum cannot be cloned”. Nature299, 802–803 (1982)

1982
[23]

Quantum Realization of the Finite Element Method

M. Deiml and D. Peterseim. “Quantum Realization of the Finite Element Method”. Math. Comp. (2025). arXiv:2403.19512

work page arXiv 2025
[24]

A polynomial quantum algorithm for approximating the

D. Aharonov, V. Jones, and Z. Landau. “A polynomial quantum algorithm for approximating the Jones polynomial”. In Proc. Thirty-Eighth Annu. ACM Symp. Theory Comput. Pages 427–436. Seattle WA USA (2006). ACM. arXiv:quant-ph/0511096

work page arXiv 2006
[25]

Amplitude estimation without phase estimation

Y. Suzuki, S. Uno, R. Raymond, T. Tanaka, T. Onodera, and N. Yamamoto. “Amplitude estimation without phase estimation”. Quantum Inf Process19, 75 (2020). arXiv:1904.10246

work page arXiv 2020
[26]

Ground-State Preparation and Energy Estimation on Early Fault-Tolerant Quantum Computers via Quantum Eigenvalue Transformation of Unitary Matrices

Y. Dong, L. Lin, and Y. Tong. “Ground-State Preparation and Energy Estimation on Early Fault-Tolerant Quantum Computers via Quantum Eigenvalue Transformation of Unitary Matrices”. PRX Quantum3, 040305 (2022)

2022
[27]

A software package for sequential quadratic programming

D. Kraft. “A software package for sequential quadratic programming”. Technical Report DFVLR-FB 88-28. Institut für Dynamik der Flugsysteme, Deutsche Forschungs- und Ver- suchsanstalt für Luft- und Raumfahrt (DFVLR), Oberpfaffenhofen (1988)

1988
[28]

Exact and efficient Lanczos method on a quantum computer

W. Kirby, M. Motta, and A. Mezzacapo. “Exact and efficient Lanczos method on a quantum computer”. Quantum7, 1018 (2023). arXiv:2208.00567

work page arXiv 2023
[29]

Efficient phase-factor evaluation in quantum signal processing

Y. Dong, X. Meng, K. B. Whaley, and L. Lin. “Efficient phase-factor evaluation in quantum signal processing”. Phys. Rev. A103, 042419 (2021). arXiv:2002.11649

work page arXiv 2021
[30]

Inverse nonlinear fast Fourier transform on SU(2) with applications to quantum signal processing

H. Ni, R. Sarkar, L. Ying, and L. Lin. “Inverse nonlinear fast Fourier transform on SU(2) with applications to quantum signal processing” (2025). arXiv:2505.12615

work page arXiv 2025
[31]

Measurement-efficient quantum Krylov subspace diagonalisation

Z. Zhang, A. Wang, X. Xu, and Y. Li. “Measurement-efficient quantum Krylov subspace diagonalisation”. Quantum8, 1438 (2024). arXiv:2301.13353

work page arXiv 2024
[32]

A Theory of Quantum Subspace Diagonalization

E. N. Epperly, L. Lin, and Y. Nakatsukasa. “A Theory of Quantum Subspace Diagonalization”. SIAM J. Matrix Anal. Appl.43, 1263–1290 (2022). arXiv:2110.07492

work page arXiv 2022
[33]

Constrained Optimal Polynomials for Quantum Linear System Solvers – Numerical Data

M. Deiml and D. Peterseim. “Constrained Optimal Polynomials for Quantum Linear System Solvers – Numerical Data”. https://zenodo.org/doi/10.5281/zenodo.19694860 (2026). A Additional data This appendix contains a full overview of the best achieved median error using each combination of method and parameters. The subscripts ∗sq and ∗′ sq indicate (8) and (9)...

work page doi:10.5281/zenodo.19694860 2026