Pith. sign in

REVIEW 3 major objections 4 minor 42 references

Optimal Quantum Eigenvalue Transformation via Linear Combinations of Hermitian Matrices

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read The paper establishes an exact identity expressing every matrix power A^m as a finite average of Chebyshev polynomials of Hermitian angular matrices, and uses it to build quantum circuits that transform functions of non-normal matrices with

desk verdict The exact Chebyshev angular projection identity is genuine and the circuit is clean, but the end-to-end optimality claim needs an explicit oracle-precision cost. read the letter →

arxiv 2607.25812 v1 pith:PV6GW7ET submitted 2026-07-28 quant-ph

Yanqiao Wang , Yixuan Liang , Hongjia Chen , Jin-Peng Liu This is my paper
classification quant-ph MSC 81P6865F60
keywords quantumeigenvaluetransformationlinearcombinationsofHermitianmatricesnon-normalChebyshevpolynomialsblockencodinggeneralizedsignalprocessingmatrixfunctionsnumericalradius
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

Every square matrix A, normal or not, can be written A=L+iH with Hermitian L and H. The paper's central discovery is that the angular Hermitian matrices X_theta = cos(theta)L + sin(theta)H carry A^m as an exact Fourier coefficient: A^m = (2/N) sum_j e^{im theta_j} T_m(X_{theta_j}) for any N>m. Because the sum is exact, with no truncation or quadrature error, polynomial functions p_d(A) reduce to a uniform average of Chebyshev transformations of Hermitian matrices. That reduction yields a quantum algorithm that block-encodes p_d(A) with Theta(d) matrix-oracle depth and the optimal post-selection repetitions O(||p_d||_infty / ||p_d(A)|psi>||), and extends to logarithms, resolvents, fractional powers, sign/ReLU, and noncircular Faber approximations.

What carries the argument

The central object is the exact Chebyshev angular projection: the identity A^m = (2/N) sum_j e^{im theta_j} T_m(X_{theta_j}), together with the observation that sup_theta ||X_theta|| = w(A), so the angular family is controlled by the numerical radius rather than the spectrum. In the quantum circuit, an angle register is prepared in a uniform superposition; each branch applies a qubitized walk whose compression is X_theta; GQSP (generalized quantum signal processing, a primitive that applies a bounded complex polynomial to a unitary signal) processes all branches coherently; uncomputing the angle register performs the root-of-unity average exactly. The angular LCU coefficients have unit l1-no

What would settle it

Take a 3x3 nilpotent Jordan block A with A^2 nonzero but A^3=0, choose m=2 and N=3, and evaluate the three-term sum (2/3) sum_{j=0}^2 e^{2i theta_j} T_2(X_{theta_j}) numerically; exactness predicts it equals A^2 to machine precision. Any deviation beyond rounding would disprove the identity. Equivalently, in a quantum implementation, measuring the success probability of the exact-power circuit on a state with known ||A^2|psi>|| would test the predicted constant 1/4.

Watch

Extended reading notes

Core claim

For any A with Cartesian decomposition A=L+iH and any integer m>=1, the paper proves A^m = (2/pi) integral_0^pi e^{im theta} T_m(X_theta) dtheta = (2/N) sum_{j=0}^{N-1} e^{im theta_j} T_m(X_{theta_j}), theta_j=pi j/N, for every N>m, where X_theta=cos(theta)L+sin(theta)H=Re(e^{-i theta}A) and T_m is the Chebyshev polynomial. The discrete identity is exact because only the leading monomial of T_m contains A^m, and the root-of-unity average isolates that Fourier mode. Applying it coefficientwise, every degree-d polynomial p_d(A) equals an equally weighted average of Hermitian matrices q_{d,theta}(X_theta), and when the numerical radius satisfies w(A)<=1 the same projection lifts to disk-algebra

Load-bearing premise

The end-to-end complexity claims assume one already possesses a block encoding of A whose error delta_A satisfies delta_A times the sum of m|c_m| is within the target error, and because that sum can grow as O(d^{3/2}) C_p, high-degree polynomials require oracle accuracy as fine as epsilon/d^{3/2}, with the cost of producing such an accurate encoding from a noisier oracle not included.

Editorial extensions

If this is right

  • A degree-d polynomial transform p_d(A) of a contraction can be block-encoded by a circuit of Theta(d) controlled matrix-oracle calls and O(d log(d+1)) angle rotations, matching the signal-query lower bound.
  • Post-selection to prepare p_d(A)|psi> requires Theta(C_p / ||p_d(A)|psi>||) coherent repetitions, and this dependence is optimal among methods valid uniformly for all contractions.
  • For exact powers, C_p=2 independent of m, so A^m is block-encoded with constant success-amplitude overhead whenever ||A^m|psi>|| = Omega(1); no 1/N factor arises from the angular average.
  • Resolvents, matrix logarithm, shifted fractional powers, sign/projector/ReLU, affine iterates, and driven ODEs all inherit near-optimal Clifford+T counts O~(d log(d/epsilon)), with approximation degree inherited from scalar Taylor or Faber approximation.
  • When the numerical radius is strictly below the operator norm, rescaling by w(A) gives an exponential-in-degree reduction in total state-preparation queries for power polynomials on a weighted-shift family.

Reading between the lines

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

  • The exactness of the root-of-unity projection suggests the same angular-sampling trick could be applied to other polynomial bases or rational approximants such as Zolotarev functions, replacing radial or contour quadrature by one-dimensional angular averages; the paper does not develop this.
  • Because the identity is purely algebraic and uses only A and A^dagger through X_theta, it may extend to unbounded operators with suitable numerical-range bounds, although the paper treats finite-dimensional matrices.
  • A testable practical corollary is that for ill-conditioned outputs the dominant cost shifts entirely to the output norm ||p_d(A)|psi>||; no circuit redesign can beat this, so the meaningful benchmark is conditioning-aware state preparation rather than query-count improvements.
  • For nilpotent or defective matrices, the construction gives Jordan-power information without diagonalization; the weighted-shift separation suggests engineered non-normal matrices are where the advantage over singular-value-based methods is largest.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper develops two 'linear combination of Hermitian matrices' (LCHM) representations for transforming functions of a generally non-normal matrix A. The first is a residue/kernel formula reducing g(A) to integrals of g(i(H+kL)), generalizing LCHS. The second, 'Weyl LCHM,' expresses A^m as an exact root-of-unity projection of Chebyshev polynomials of the angular Hermitian matrices X_θ = Re(e^{-iθ}A). The central identity is A^m = (2/N) Σ_{j=0}^{N-1} e^{imθ_j} T_m(X_{θ_j}) for any N>m, with no truncation or quadrature error. The authors use this identity to construct a coherent quantum circuit, based on GQSP and qubitization, that block-encodes p_d(A) in Θ(d) oracle depth and with post-selection overhead claimed to be optimal. The framework is extended to disk-algebra functions, Faber approximation, and a wide set of matrix function examples (exponentials, resolvents, logarithms, fractional powers, sign/ReLU, driven ODEs). A numerical-radius-rescaled variant is claimed to give an exponential improvement over prior d-regular GQSP on some weighted-shift families.

Significance. If the resource claims hold, this is a substantial contribution to quantum eigenvalue transformation. The Weyl identity is elegant and noncommutative; the proof via Fourier-mode counting is convincing, and the root-of-unity discretization being exact rather than a quadrature approximation is a real advantage. The coherent GQSP construction appears sound and the error analysis is detailed, including perturbation and Clifford+T costs. The paper also gives explicit, falsifiable complexity bounds for several concrete matrix functions. The use of the numerical radius through sup_θ ||X_θ|| = w(A) is insightful and leads to a concrete separation example. These are strengths that go beyond a purely incremental paper.

major comments (3)
  1. [§2.1, Eq. (2.2) and Theorem 2.1] The printed kernel f_β(z) = (1/2π) exp(2β - (1+iz)^β)/(1-iz) has residue i e^{2β - 2^β}/(2π) at z=-i, not i/(2π). Equation (2.7) therefore gives e^{2β-2^β}g(A), not g(A), unless '2β' is a rendering of 2^β. As written, Theorem 2.1 and its use in the LCHS specialization (5.1) are incorrect. Please correct the notation to exp(2^β - (1+iz)^β) throughout.
  2. [§4.2, Theorem 4.5, Eq. (4.42); Appendix D, Eq. (D.3)] The end-to-end resource claim does not account for the cost of obtaining the oracle accuracy required by the error bound. The term (δ_A/2)Σ m|c_m| in (4.42), together with the Parseval-type bound (D.3), implies that for a general degree-d polynomial one needs δ_A = O(ε/(C_p d^{3/2})) to reach unnormalized error ε. The counts (4.37) and Table 2 count calls to a U_A that is already this accurate. If the raw input oracle is less accurate, reducing δ_A to the required value costs additional queries, and that cost is omitted. The abstract's 'Θ(d) circuit depth' should be qualified as holding for a supplied block encoding of accuracy δ_A within the target error, or an explicit discussion of the precision-amplification overhead should be added.
  3. [Appendix A, Eqs. (A.7)-(A.16)] The claimed exponential advantage from numerical-radius rescaling relies on uniform singular-value amplification of the map x↦x/s on the Hermitian matrices X_θ. However, the error analysis for this inner amplification and its propagation through the subsequent degree-d GQSP pipeline and amplitude amplification is only sketched, hidden in the eO notation. Since this exponential separation is a headline contribution (item (iv) in the Introduction and again in the Conclusion), it needs a rigorous theorem analogous to Theorem 4.5, specifying the polynomial degree, approximation error, and total query count including the amplification step.
minor comments (4)
  1. [Eq. (5.1) and abstract] The same '2β' rendering issue appears in the LCHS formula (5.1) and elsewhere; please ensure the intended 2^β notation is used consistently.
  2. [Abstract and Corollary 4.6] The phrase 'post-selection repetitions' in the abstract could be misread as independent postselection attempts, which scale as Θ(C_p^2/||p_d(A)|ψ⟩||^2). The O(C_p/||p_d(A)|ψ⟩||) result is for coherent amplitude amplification. Please use 'amplitude-amplified repetitions' or explicitly state the distinction.
  3. [§7.1, Table 3 caption] There is a typo: 'under the common assumption ∥A∥ ≤ 1 and and given a ...' should read 'and given a ...'.
  4. [Acknowledgments] The paragraph beginning 'Human authors came up with the LCHM formulas...' is nonstandard and should be removed or rewritten; it does not belong in a scientific paper's acknowledgments.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central A^m identity is derived from the Chebyshev/Fourier structure, not assumed.

full rationale

The core claim, Theorem 3.2, is proven in the text: the leading coefficient of T_m gives the e^{-imθ} mode of T_m(X_θ) as A^m/2, and the root-of-unity sum projects exactly because N>m. The quantum implementation then follows from this identity using standard GQSP/qubitization; Eq. (4.35) is an exact algebraic consequence, and the error bound (4.42) is derived by telescoping estimates in Appendix D, not by fitting. The optimality lower bounds use independent scalar-contraction and sequential-signal arguments, not the algorithm itself. Self-citations such as [19] (LCHS) and [32] (sign embedding) are used as context or comparisons and are not load-bearing; the key external primitives—GQSP, qubitization, von Neumann's inequality, Crouzeix–Palencia, Okubo–Ando—are cited independently, and the qubitized Chebyshev identity is re-proved in Section C. The skeptic's concern about the cost of supplying a sufficiently accurate block encoding δ_A is a resource-accounting caveat (correctness/completeness risk), not a circular step, since no prediction reduces to an input by construction.

Assumptions & free parameters 2 free parameters · 8 assumptions · 0 invented entities

No new physical entities or fitted constants are introduced. The angular Hermitian pencil X_θ and the lifted scalar F_g are notational/mathematical constructions, not postulated entities with independent evidence requirements. The main uncharged ingredients are the external QSP/qubitization primitives and the metric information (∥A∥, w(A), block-encoding precision) assumed as input.

free parameters (2)
  • kernel exponent β = any 0<β<1 (e.g., β=1/2)
    Vanilla LCHM kernel fβ uses a tunable β to control subexponential decay; the theorem is stated for all β in (0,1), so it is a design choice rather than a fitted value. The printed expression has a residue-normalization ambiguity at Eq. (2.2)-(2.3).
  • numerical-radius rescale factor s = w(A)(1+1/d) in the weighted-shift analysis
    Chosen by hand in Section A.1-A.2 to balance the degree-k coefficient damping s^k against the inner amplification cost 1/(s-w(A)); correctness holds for any s∈(w(A),1), but the exponential advantage is optimized by this choice.
assumptions (8)
  • standard math Okubo–Ando inequality: if w(A)≤1, then ||r(A)||≤2||r||∞,D for polynomials r
    Used to define the disk-algebra functional calculus and to transfer scalar approximation to operator approximation (Section 3.2, Definition 3.4, Theorem 3.5).
  • standard math Crouzeix–Palencia (1+√2)-spectral-set bound on the numerical range
    Used in the proof of vanilla LCHM (Theorem 2.1) and to transfer Faber approximation errors (Theorem 6.1).
  • standard math Generalized quantum signal processing: any degree-d complex polynomial P with |P|≤1 on the unit circle is implementable with O(d) calls to a unitary signal
    Core implementation primitive in Theorem 4.5; imported from Motlagh–Wiebe (Ref. [16]).
  • standard math Qubitization walk identity ΠW^mΠ = T_m(X)Π for an involutive direct dilation
    Proved in Appendix C by reduction to a 2×2 rotation; used in Eq. (4.28) to realize Chebyshev transforms of the angular Hermitian matrices.
  • standard math von Neumann's inequality and the scalar contraction realization of ∥p∥∞,D
    Used in Proposition 4.8 to establish the optimal post-selection lower bound in the uniform contraction model.
  • domain assumption A unit-normalized block encoding of a contraction A, with coherent access to controlled U_A, U_A†, and reflections
    The quantum algorithm in Section 4 assumes a (1,a,δ_A) block encoding with ∥A∥≤1 and coherent oracle access; without this the circuit construction and depth count do not apply.
  • domain assumption A certified numerical-radius bound w(A)<s<1 for the numerical-radius rescaling
    The exponential advantage of Section A depends on a strict numerical-radius gap and on a certifiable s; if w(A)=∥A∥, the construction reduces to the constant-factor comparison.
  • domain assumption The function g lies in the disk algebra D(D) or is analytic in a neighborhood of the relevant numerical-range enclosure
    The disk-algebra functional calculus and the Faber–Weyl extension require holomorphy on the relevant sets (Theorems 3.5, 4.1, 6.1).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimal Quantum Eigenvalue Transformation via Linear Combinations of Hermitian Matrices." pith.science (2026). https://pith.science/paper/PV6GW7ET

@misc{pith2026260725812,
  author       = {Pith},
  title        = {Pith review of: Optimal Quantum Eigenvalue Transformation via Linear Combinations of Hermitian Matrices},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PV6GW7ET}},
  note         = {Machine review of arXiv:2607.25812}
}
abstract

We discover two complementary linear-combination-of-Hermitian-matrices (LCHM) formulations to achieve a general non-normal matrix eigenvalue transformation $g(A)$. Firstly, for $A=L+\mathrm{i} H$ with Hermitian $L$ and $H$, the vanilla LCHM formula represents $g(A)$ as a kernel integral of $g(\mathrm{i}(H+kL))$, and it contains linear-combination-of-Hamiltonian-simulation (LCHS) [An, Liu, Lin, Phys. Rev. Lett. 2023] as the special case for matrix exponentials. Secondly, for the angular Hermitian $X_\theta = \cos\theta L+\sin\theta H$, the Weyl LCHM formula expresses $g(A)$ via integrating $g(\mathrm{e}^{\mathrm{i}\theta} (X_\theta\pm\mathrm{i}(I-X_\theta^2)^{1/2}))$. For the matrix power $g(A)=A^m$, the Fourier projection of Weyl LCHM gives \[ A^m=\frac{2}{\pi}\int_0^\pi \text{e}^{\text{i} m\theta}T_m(X_\theta) \text{d}\theta = \frac{2}{N}\sum_{j=0}^{N-1} \text{e}^{\text{i} m\theta_j}T_m(X_{\theta_j}),\quad\theta_j=\frac{\pi j}{N},\quad \text{for every } N>m \] with Chebyshev polynomial of Hermitian $T_m(X_\theta)$ and $N$ samples. The discrete formula is exact, introduces no truncation and angular quadrature error, and offers $\mathcal{O}(1)$ post-selection weights. LCHM formulas lead to new quantum eigenvalue transformation (QET) algorithms. For a degree-$d$ polynomial $p_d(A)$ on $|\psi\rangle$, our QET algorithm can achieve optimal $\Theta(d)$ circuit depth and optimal $\mathcal{O}(||p_d||_{\infty}/||p_d(A)|\psi\rangle||)$ post-selection repetitions. LCHM-based QETs unify various quantum linear algebraic problems with near-optimal $\mathcal{\widetilde O}(d\log(d/\epsilon))$ Clifford$+T$ gates, including driven ODEs (reduced to standard LCHS), iterative methods, resolvents, $\log(I+A)$, $(\lambda I+A)^\nu$, Sign and ReLU transforms, and Faber approximation on noncircular domains.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

42 extracted references · 3 canonical work pages

  1. [1]

    Higham.Functions of Matrices: Theory and Computation

    Nicholas J. Higham.Functions of Matrices: Theory and Computation. SIAM, Philadelphia, 2008.doi:10.1137/1.9780898717778

    doi:10.1137/1.9780898717778 2008

  2. [2]

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

    Andr´ as Gily´ en, Yuan Su, Guang Hao Low, and Nathan Wiebe. Quantum singular value transformation and beyond: Exponential improvements for quantum matrix arithmetics. In Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing, pages 193–204, 2019.doi:10.1145/3313276.3316366

    arXiv 2019

  3. [3]

    Universal quantum simulators.Science, 273(5278):1073–1078, 1996.doi:10

    Seth Lloyd. Universal quantum simulators.Science, 273(5278):1073–1078, 1996.doi:10. 1126/science.273.5278.1073

    arXiv 1996

  4. [4]

    Childs and Nathan Wiebe

    Andrew M. Childs and Nathan Wiebe. Hamiltonian simulation using linear combinations of unitary operations.Quantum Information and Computation, 12(11–12):901–924, 2012.doi: 10.26421/QIC12.11-12-1

    doi:10.26421/qic12.11-12-1 2012

  5. [5]

    Berry, Andrew M

    Dominic W. Berry, Andrew M. Childs, Richard Cleve, Robin Kothari, and Rolando D. Somma. Simulating Hamiltonian dynamics with a truncated Taylor series.Physical Review Letters, 114:090502, 2015.doi:10.1103/PhysRevLett.114.090502

    doi:10.1103/physrevlett.114.090502 2015

  6. [6]

    Guang Hao Low and Isaac L. Chuang. Optimal Hamiltonian simulation by quantum signal pro- cessing.Physical Review Letters, 118:010501, 2017.doi:10.1103/PhysRevLett.118.010501. 48

    doi:10.1103/physrevlett.118.010501 2017

  7. [7]

    Guang Hao Low and Isaac L. Chuang. Hamiltonian simulation by qubitization.Quantum, 3:163, 2019.doi:10.22331/q-2019-07-12-163

    doi:10.22331/q-2019-07-12-163 2019

  8. [8]

    Harrow, Avinatan Hassidim, and Seth Lloyd

    Aram W. Harrow, Avinatan Hassidim, and Seth Lloyd. Quantum algorithm for linear systems of equations.Physical Review Letters, 103:150502, 2009.doi:10.1103/PhysRevLett.103. 150502

    doi:10.1103/physrevlett.103 2009

Show all 42 references
  1. [9]

    Childs, Robin Kothari, and Rolando D

    Andrew M. Childs, Robin Kothari, and Rolando D. Somma. Quantum algorithm for systems of linear equations with exponentially improved dependence on precision.SIAM Journal on Computing, 46(6):1920–1950, 2017.doi:10.1137/16M1087072

    1920 doi

  2. [10]

    Pedro C. S. Costa, Dong An, Yuval R. Sanders, Yuan Su, Ryan Babbush, and Dominic W. Berry. Optimal scaling quantum linear-systems solver via discrete adiabatic theorem.PRX Quantum, 3:040303, 2022.doi:10.1103/PRXQuantum.3.040303

    2022 doi

  3. [11]

    Dominic W. Berry. High-order quantum algorithm for solving linear differential equa- tions.Journal of Physics A: Mathematical and Theoretical, 47:105301, 2014.doi:10.1088/ 1751-8113/47/10/105301

    2014

  4. [12]

    Berry, Andrew M

    Dominic W. Berry, Andrew M. Childs, Aaron Ostrander, and Guoming Wang. Quan- tum algorithm for linear differential equations with exponentially improved dependence on precision.Communications in Mathematical Physics, 356:1057–1081, 2017.doi:10.1007/ s00220-017-3002-y

    2017

  5. [13]

    Childs and Jin-Peng Liu

    Andrew M. Childs and Jin-Peng Liu. Quantum spectral methods for differential equa- tions.Communications in Mathematical Physics, 375:1427–1457, 2020.doi:10.1007/ s00220-020-03699-z

    2020

  6. [14]

    Improved quantum algorithms for linear and nonlinear differential equations

    Hari Krovi. Improved quantum algorithms for linear and nonlinear differential equations. Quantum, 7:913, 2023.doi:10.22331/q-2023-02-02-913

    2023 doi

  7. [15]

    Time-marching based quantum solvers for time-dependent linear differential equations.Quantum, 7:955, 2023.doi:10.22331/q-2023-03-20-955

    Di Fang, Lin Lin, and Yu Tong. Time-marching based quantum solvers for time-dependent linear differential equations.Quantum, 7:955, 2023.doi:10.22331/q-2023-03-20-955

    2023 doi

  8. [16]

    Generalized quantum signal processing.PRX Quantum, 5:020368, 2024.doi:10.1103/PRXQuantum.5.020368

    Danial Motlagh and Nathan Wiebe. Generalized quantum signal processing.PRX Quantum, 5:020368, 2024.doi:10.1103/PRXQuantum.5.020368

    2024 doi

  9. [17]

    Quantum eigenvalue processing.SIAM Journal on Computing, 55(1):135–215, 2026.doi:10.1137/24M1689363

    Guang Hao Low and Yuan Su. Quantum eigenvalue processing.SIAM Journal on Computing, 55(1):135–215, 2026.doi:10.1137/24M1689363

    2026 doi

  10. [18]

    Quantum eigenvalue transformations for arbitrary matrices, 2026.arXiv:2604.19688

    Xabier Guti´ errez, Lorenzo Laneve, and Mikel Sanz. Quantum eigenvalue transformations for arbitrary matrices, 2026.arXiv:2604.19688

    2026 arXiv

  11. [19]

    Dong An, Jin-Peng Liu, and Lin Lin. Linear combination of Hamiltonian simulation for nonunitary dynamics with optimal state preparation cost.Physical Review Letters, 131:150603, 2023.doi:10.1103/PhysRevLett.131.150603. 49

    2023 doi

  12. [20]

    Childs, and Lin Lin

    Dong An, Andrew M. Childs, and Lin Lin. Quantum algorithm for linear non-unitary dynamics with near-optimal dependence on all parameters.Communications in Mathematical Physics, 407:19, 2026.doi:10.1007/s00220-025-05509-w

    2026 doi

  13. [21]

    Guang Hao Low and Rolando D. Somma. Optimal quantum simulation of linear non-unitary dynamics, 2025.arXiv:2508.19238

    2025 arXiv

  14. [22]

    Childs, Lin Lin, and Lexing Ying

    Dong An, Andrew M. Childs, Lin Lin, and Lexing Ying. Laplace transform–based quantum eigenvalue transformation via linear combination of Hamiltonian simulation.SIAM Journal on Computing, 55(2):376–409, 2026.doi:10.1137/24M1720160

    2026 doi

  15. [23]

    Fourier transform–based linear combination of Hamiltonian simula- tion, 2025.arXiv:2508.19596

    Xi Huang and Dong An. Fourier transform–based linear combination of Hamiltonian simula- tion, 2025.arXiv:2508.19596

    2025 arXiv

  16. [24]

    Quantum eigenvalue transformation via linear combination of Hamiltonian simulation: A Weyl calculus approach, 2026.arXiv:2606.29848

    Hongkang Ni and Lexing Ying. Quantum eigenvalue transformation via linear combination of Hamiltonian simulation: A Weyl calculus approach, 2026.arXiv:2606.29848

    2026 arXiv

  17. [25]

    Quantum simulation of partial differential equations via Schr¨ odingerization.Physical Review Letters, 133:230602, 2024.doi:10.1103/PhysRevLett

    Shi Jin, Nana Liu, and Yue Yu. Quantum simulation of partial differential equations via Schr¨ odingerization.Physical Review Letters, 133:230602, 2024.doi:10.1103/PhysRevLett. 133.230602

    2024 doi

  18. [26]

    Quantum simulation of partial differential equations: Appli- cations and detailed analysis.Physical Review A, 108:032603, 2023.doi:10.1103/PhysRevA

    Shi Jin, Nana Liu, and Yue Yu. Quantum simulation of partial differential equations: Appli- cations and detailed analysis.Physical Review A, 108:032603, 2023.doi:10.1103/PhysRevA. 108.032603

    2023 doi

  19. [27]

    Quantum simulation of discrete linear dynamical systems and simple iterative methods in linear algebra.Proceedings of the Royal Society A, 480:20230370, 2024

    Shi Jin and Nana Liu. Quantum simulation of discrete linear dynamical systems and simple iterative methods in linear algebra.Proceedings of the Royal Society A, 480:20230370, 2024. doi:10.1098/rspa.2023.0370

    2024

  20. [28]

    Higham, and Lloyd N

    Nicholas Hale, Nicholas J. Higham, and Lloyd N. Trefethen. ComputingAα, log(A), and related matrix functions by contour integrals.SIAM Journal on Numerical Analysis, 46(5):2505–2523, 2008.doi:10.1137/070700607

    2008 doi

  21. [29]

    Souichi Takahira, Asuka Ohashi, Tomohiro Sogabe, and Tsuyoshi S. Usuda. Quantum algo- rithm for matrix functions by Cauchy’s integral formula.Quantum Information and Compu- tation, 20(1–2):14–36, 2020.doi:10.26421/QIC20.1-2-2

    2020 doi

  22. [30]

    Souichi Takahira, Asuka Ohashi, Tomohiro Sogabe, and Tsuyoshi S. Usuda. Quantum algo- rithms based on the block-encoding framework for matrix functions by contour integrals.Quan- tum Information and Computation, 22(11–12):965–979, 2022.doi:10.26421/QIC22.11-12-4

    2022 doi

  23. [31]

    Contour-integral based quantum eigenvalue transformation: Anal- ysis and applications, 2026.arXiv:2601.11959

    Shan Jiang and Dong An. Contour-integral based quantum eigenvalue transformation: Anal- ysis and applications, 2026.arXiv:2601.11959

    2026

  24. [32]

    Sign embedding quantum algorithms for matrix equations and matrix functions, 2026.arXiv:2604.25333

    Yanqiao Wang and Jin-Peng Liu. Sign embedding quantum algorithms for matrix equations and matrix functions, 2026.arXiv:2604.25333. 50

    2026 arXiv

  25. [33]

    The computation of functions of matrices by truncated Faber series.Numerical Functional Analysis and Optimization, 22(5–6):697–719, 2001.doi:10

    Igor Moret and Paolo Novati. The computation of functions of matrices by truncated Faber series.Numerical Functional Analysis and Optimization, 22(5–6):697–719, 2001.doi:10. 1081/NFA-100105314

    2001

  26. [34]

    Trefethen.Approximation Theory and Approximation Practice

    Lloyd N. Trefethen.Approximation Theory and Approximation Practice. SIAM, Philadelphia, extended edition, 2019.doi:10.1137/1.9781611975949

    2019 doi

  27. [35]

    The numerical range is a (1 + √ 2)-spectral set.SIAM Journal on Matrix Analysis and Applications, 38(2):649–655, 2017.doi:10.1137/17M1116672

    Michel Crouzeix and C´ esar Palencia. The numerical range is a (1 + √ 2)-spectral set.SIAM Journal on Matrix Analysis and Applications, 38(2):649–655, 2017.doi:10.1137/17M1116672

    2017 doi

  28. [36]

    Constants related to operators of classC ρ.Manuscripta Mathematica, 16(4):385–394, 1975.doi:10.1007/BF01323467

    Kazuyoshi Okubo and Tsuyoshi Ando. Constants related to operators of classC ρ.Manuscripta Mathematica, 16(4):385–394, 1975.doi:10.1007/BF01323467

    1975 doi

  29. [37]

    Springer, New York, second edition, 2010.doi:10.1007/ 978-1-4419-6094-8

    B´ ela Sz.-Nagy, Ciprian Foia¸ s, Hari Bercovici, and L´ aszl´ o K´ erchy.Harmonic Analysis of Operators on Hilbert Space. Springer, New York, second edition, 2010.doi:10.1007/ 978-1-4419-6094-8

    2010

  30. [38]

    Quantum amplitude amplifi- cation and estimation

    Gilles Brassard, Peter Høyer, Michele Mosca, and Alain Tapp. Quantum amplitude amplifi- cation and estimation. In Samuel J. Lomonaco, Jr. and Howard E. Brandt, editors,Quantum Computation and Information, volume 305 ofContemporary Mathematics, pages 53–74. Amer- ican Mathemati...

    2002 doi

  31. [39]

    Grover’s quantum searching algorithm is optimal.Physical Review A, 60:2746– 2751, 1999.doi:10.1103/PhysRevA.60.2746

    Christof Zalka. Grover’s quantum searching algorithm is optimal.Physical Review A, 60:2746– 2751, 1999.doi:10.1103/PhysRevA.60.2746

    1999 doi

  32. [40]

    Quantum algorithms for solving generalized linear systems via momentum accelerated gradient and Schr¨ odingerization, 2025

    Qitong Hu, Xiaoyang He, Shi Jin, and Xiao-Dong Zhang. Quantum algorithms for solving generalized linear systems via momentum accelerated gradient and Schr¨ odingerization, 2025. arXiv:2509.16576

    2025

  33. [41]

    Yuji Nakatsukasa and Roland W. Freund. Computing fundamental matrix decompositions accurately via the matrix sign function in two iterations: The power of Zolotarev’s functions. SIAM Review, 58(3):461–493, 2016.doi:10.1137/140990334

    2016 doi

  34. [42]

    Kenney and Alan J

    Charles S. Kenney and Alan J. Laub. The matrix sign function.IEEE Transactions on Automatic Control, 40(8):1330–1348, 1995.doi:10.1109/9.402226. A Numerical-radius-rescaled Weyl LCHM Numerical-radius rescaling converts a uniform gapw(A)<1 into degree-dependent coefficient damp...

    1995 doi

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.