Pith. sign in

REVIEW 3 major objections 4 minor 35 references

This paper constructs explicit block-encodings for discretized biharmonic equations with periodic, simply supported, and Dirichlet–Neumann boundary conditions, and proves that QSVT–VTAA linear-system solvers then prepare solution states wit

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

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

T0 review · deepseek-v4-flash

2026-08-01 04:52 UTC pith:ODKKQB5I

load-bearing objection The periodic and simply supported sections deliver clean block-encodings with O(N^2) condition numbers; the direct Dirichlet–Neumann scheme has a genuinely new construction but rests on an unproven stability inequality in Lemma 5.2, so the paper needs a serious revision, not a desk reject. the 3 major comments →

arxiv 2607.22396 v1 pith:ODKKQB5I submitted 2026-07-24 quant-ph cs.NAmath.NA

Explicit block-encodings for biharmonic boundary-value problems

Chuwen Ma , Zihao Tang This is my paper
classification quant-ph cs.NAmath.NA MSC 65N0665N3581P68
keywords biharmonic equationblock-encodingquantum linear-system algorithmsQSVTVTAAboundary conditionsfinite differencescondition number
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 paper's central claim is that the biharmonic equation — a fourth-order PDE — can be solved by quantum linear-system algorithms once explicit block-encodings are written down for each common boundary condition, and that the augmented two-Laplacian formulations for periodic and simply supported cases have condition number O(N^2) instead of the naive O(N^4). For Dirichlet–Neumann data, the paper introduces a boundary-corrected finite-difference discretization, proves a mesh-independent stability bound, block-encodes the resulting nonsymmetric matrix with optimal normalization, and shows the condition number is O(N^4); a coupled-Laplace variant moves boundary values of the auxiliary variable into extra unknowns whose empirical condition number grows like N^3 in 2D. A sympathetic reader would care because these constructions turn abstract QSVT machinery into concrete circuits with controlled normalization and gate cost, yielding explicit fixed-grid and end-to-end complexity estimates and small-scale numerical verification.

Core claim

The core discovery is that the fourth-order biharmonic operator, after discretization, need not be inverted as a squared Laplacian: rewriting it as an augmented block-triangular system of two second-order operators brings the condition number down to that of a Poisson system when Fourier or sine diagonalization applies. For Dirichlet–Neumann boundary conditions, where v = -Δu has unknown boundary values, the paper shows a consistent one-sided boundary closure keeps global second-order convergence and allows an exact norm-matched block-encoding of the nonsymmetric boundary correction, giving a direct O(N^4)-conditioned system; an alternative coupled-Laplace system with boundary unknowns admit

What carries the argument

The central object is the block-encoding, a unitary whose top-left block equals a scaled matrix of interest; the paper builds them via LCU from shift-and-reflection encodings of the 1D Dirichlet Laplacian, from diagonal (Fourier or sine) encodings, and from a norm-matched two-qubit state that encodes the boundary-row correction. The augmented Poisson matrices P(B)=[[B,-I],[0,B]] acquire block-encodings with normalization ‖B‖+1 by a three-term LCU, and the QSVT–VTAA solver converts each block-encoding into nearly-linear-in-κ state preparation.

Load-bearing premise

The load-bearing premise is that the 4-by-4 boundary-layer computation in Lemma 5.2 really implies a mesh-independent stability bound for the whole boundary-corrected Dirichlet–Neumann matrix; the proof asserts the 'combining' step without a detailed argument.

What would settle it

Compute, for N=10,20,…,1000, the ratio λ_min(H(T_N^2+M_0)) / λ_min(T_N^2); if the ratio tends to zero as N grows, Lemma 5.2's c>0 independent of N does not exist, invalidating the direct scheme's stability. Alternatively, run the three-dimensional test from Section 7.3 at higher N and check whether the L2 error order drops below 2.

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

If this is right

  • For periodic and simply supported biharmonic problems, quantum state-preparation costs scale as Õ(d^2 N^2) in the smooth regime or Õ(d^3 N^4) in the worst case, with condition number O(N^2) rather than O(N^4).
  • The direct Dirichlet–Neumann discretization is second-order convergent and mesh-independently stable, with an explicit exact block-encoding and fixed-grid gate complexity Õ(d N^4).
  • The coupled-Laplace formulation yields an explicit block-encoding with normalization O(d N^2) and gate cost O(d log N); its quantum complexity is governed by κ(K), numerically ~N^3 in 2D.
  • The analysis separates discretization error from quantum state-preparation error; choosing N = O(ε^{-1/2}) gives end-to-end complexity Õ(d ε^{-2}) for the direct scheme.
  • Small-scale classical circuit simulations confirm that the constructed circuits realize the intended discrete linear solves for periodic, simply supported, and Dirichlet–Neumann problems.

Where Pith is reading between the lines

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

  • The condition-number improvement from O(N^4) to O(N^2) in the augmented formulations suggests that any fourth-order elliptic problem that can be split into second-order systems may inherit the cheaper scaling; a testable extension is applying the same augmented-block-encoding trick to mixed formulations of Stokes or elasticity problems.
  • The mesh-independent stability lemma is the only step whose proof is asserted rather than fully written; if the 'combining' step fails, the direct Dirichlet–Neumann scheme loses both its convergence and its block-encoding norm bound, so a complete proof or a counterexample would settle the matter.
  • The coupled-Laplace empirical scaling κ(K)≈N^3, if it holds in higher dimensions, would give a better exponent than the direct scheme's O(N^4) worst case, and suggests that a preconditioned or spectral-collocation variant might reduce the exponent further.
  • The paper's input/output model assumes efficient preparation of amplitude-encoded right-hand sides; the practical advantage over classical solvers depends on that model, since classical algorithms produce full vectors while the quantum output is a state.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. This paper proposes explicit block-encodings and QSVT--VTAA quantum linear-system algorithms for the biharmonic equation under four boundary treatments: periodic, simply supported, direct Dirichlet--Neumann, and coupled-Laplace splitting. For the first two cases, Fourier/sine diagonalization leads to augmented second-order systems with condition number O(dN^2) or O(N^2) and gate complexity \tilde O(N^2) per solution. For the direct Dirichlet--Neumann discretization, the paper introduces a one-sided boundary-corrected finite-difference scheme, claims mesh-independent stability H((L_h)^2+M) \succeq cI, constructs an exact norm-matched block-encoding with normalization O(N^4), and derives \kappa(\tilde A)=O(N^4). A coupled-Laplace formulation with boundary unknowns is also block-encoded, with complexity expressed in terms of the empirically observed \kappa(K)\approx N^3. Numerical experiments on small grids verify the circuit implementations and report second-order convergence for the Dirichlet--Neumann schemes.

Significance. If the main stability lemma were fully proved, the paper would provide a genuinely explicit, gate-level treatment of fourth-order elliptic boundary-value problems in the QSVT framework, including normalization factors and condition-number scalings. The periodic and simply supported sections are self-contained and rigorous, and the explicit circuits for the nonsymmetric boundary-corrected matrix and for the boundary operators are valuable. The paper also separates discretization error from state-preparation error, ships reproducible code, and validates the circuits against classical solves. The main weakness is that the Dirichlet--Neumann direct scheme's convergence, normalization overhead, and condition-number estimate all rest on a stability proof that is currently asserted rather than demonstrated.

major comments (3)
  1. [§5.2, Lemma 5.2] The proof verifies the 4×4 matrix E = diag(0,1,1,1)+B and its spectrum, then states that 'combining the two boundary layers with the positive interior contribution gives H(T_N^2+M_0) ⪰ cT_N^2.' The combination step is not shown. One must exhibit a decomposition of the full vector d = -T_N v into left, right, and interior parts, bound the left/right boundary quadratic forms by neighboring components of d, and handle the artificial component d0 = 2v1 that is not part of ∥d∥. The positive spectrum of E makes the conclusion plausible, but the global inequality is the load-bearing input for Theorem 5.3, Lemma 5.4, and Theorem 5.5. Please supply a complete proof, including the simultaneous treatment of both boundaries and the case of small N where the layers may overlap.
  2. [§5.2, Theorem 5.3 and §5.3, Lemma 5.4] The extension to d dimensions is dispatched as 'by tensor-product summation' without details. In the d-dimensional operator H(Ã) = ∑ℓ (L_ℓ² + H(M_ℓ)) + 2∑_{ℓ<j} L_ℓ L_j, the cross terms are positive semidefinite only because the L_ℓ commute; this should be stated and proved. More importantly, the paper should justify that the 1D estimate yields H(Ã) ⪰ c(L^{(d)}_h)² in all d, including any edge/corner interactions of the boundary corrections. In the proof of Theorem 5.5, the displayed lower bound σ_min(Ã) ≥ C_d λ_min²(L_h^{(1)}) + d(d-1)/2 λ_min²(L_h^{(1)}) does not follow transparently from Lemmas 5.1–5.2; the correct constant should come from λ_min(L^{(d)}_h) = d λ_min(L_h^{(1)}). Please rewrite this argument explicitly.
  3. [§5.3, Lemma 5.4 and §5.1, Eq. (23)] The normalization α_M = ∥M∥ = √21433/12 (N+1)^4 corresponds to h = 1/(N+1), i.e. a unit interval, whereas Eq. (23) states Ω=[-1,1]^d and §5.1 gives a mesh-size expression written as 2/(2n+1) that is inconsistent with either convention. The numerical test in §7.3 uses the unit interval. This ambiguity affects the exact constant in α_Ã and the claimed 'norm-matched' property of the block-encoding. Please reconcile the domain, the mesh size, and the normalization constants.
minor comments (4)
  1. [§2.2, Lemma 2.3] The displayed gate complexity contains a duplicated log κ(A) factor; the intended expression appears to be O(β_A κ(A) log κ(A) T_A log(1/ε)).
  2. [§5.3, after Lemma 5.4] The line '∥Ã∥₂ = ∥Ã∥₂ + ∥Ã^T∥/2 ≥ ∥H(Ã)∥₂' is misprinted; the correct inequality is ∥Ã∥₂ ≥ ∥(Ã+Ã^T)/2∥₂.
  3. [§6.3 and §7.4] The empirical scaling κ(K)≈N³ is used in the complexity discussion of Section 6. Please state explicitly in Section 6.3 that this is a numerical observation and not a theorem; Table 1 already notes this, but the main text should avoid implying a proven scaling.
  4. [Throughout] Typos and formatting: 'QSVTmethodcombinedwithvariabletime' in §2; 'thequantum discrete sine transform' in Theorem 4.2; Figure 13 axis labels are broken ('1.00 0.75...'). Also, the notation N is used both for the 1D resolution and for the total number of scalar unknowns N=N^d; although the Notation section defines this, the repeated use in Table 1 could confuse readers.

Circularity Check

0 steps flagged

No significant circularity: block-encodings are explicit circuit constructions with normalization derived from matrix norms; the only flagged gap (Lemma 5.2) is an unproven 'combining' step, not a definitional reduction. Self-citation [19] is background only.

full rationale

The derivation chain is self-contained. Periodic (Sec. 3) and simply supported (Sec. 4) block-encodings are constructed from explicit diagonal forms (Λ_N, D^(1)) with normalization factors computed from matrix spectra (Lemmas 3.2, 4.1); condition numbers κ(P_N)=Θ(dN^2), κ(D_N)=O(N^2) follow from eigenvalue estimates, and the QSVT–VTAA cost is quoted from the external result [20] (Lemma 2.3). No parameter is fitted to data: ρ_N is not an unknown fitted constant, and the paper explicitly labels κ(K) in the coupled formulation as a numerical estimate (Table 1, Sec. 6.3). The direct Dirichlet–Neumann construction is an explicit block-encoding of M via the constant-size state |u_M⟩ defined from the row entries; this is construction, not circularity. The only structurally weak point is Lemma 5.2, where after computing the boundary-layer matrix spectrum the proof asserts 'Combining the two boundary layers with the positive interior contribution gives H(T_N^2+M_0) ⪰ c T_N^2' without displaying the cross-term inequality; and Theorem 5.3's d-dimensional extension via 'tensor-product summation' is stated, not proved. That is an incompleteness/correctness risk, not a circular reduction: the asserted inequality is not assumed as the input and does not coincide by definition with the lemma's conclusion. The single self-citation [19] (shared author Ma) appears only in the introduction as background on Schrödingerization and is not used to justify any main theorem. No equivalence between a predicted quantity and a fitted parameter, and no uniqueness/ansatz imported solely from the authors' prior work, is present.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 0 invented entities

The paper introduces no new physical entities. The additional boundary unknowns v_B in the coupled system are mathematical variables, not new postulates. The main assumptions are the standard input/output model, the null-space convention, and two unproven or partially proven stability statements (Lemma 5.2 and its d-dimensional extension), plus an empirical condition-number scaling.

axioms (5)
  • domain assumption Assumption 1.1: the normalized quantum states for the right-hand side and boundary data are efficiently preparable, and their preparation costs are excluded.
    This is the input/output model used in every complexity statement. If this assumption fails, the claimed gate complexities do not apply.
  • domain assumption Assumption 1.2: when the operator has a null space (periodic case), the problem is restricted to the orthogonal complement; the solution is chosen with zero mean.
    This is a standard treatment for periodic problems and is stated explicitly in Section 1.
  • domain assumption Lemma 5.2: H(T_N^2+M_0) ⪰ c T_N^2 with c>0 independent of N for the one-dimensional boundary-corrected matrix.
    The text attempts a proof but jumps from a 4x4 boundary-layer eigenvalue computation to the global operator inequality. The claimed inequality is load-bearing for Theorem 5.3 and the condition-number bound, but the proof as written is incomplete.
  • domain assumption Theorem 5.3: the d-dimensional stability of the boundary-corrected discretization follows by 'tensor-product summation' from the one-dimensional result.
    The extension to d dimensions is asserted rather than demonstrated. It is needed for the convergence and complexity claims in the Dirichlet–Neumann case.
  • domain assumption Empirical scaling κ(K)≈N^3 for the coupled-Laplace system in two dimensions (Section 7.4).
    The complexity of the coupled formulation is expressed in terms of κ(K), but no rigorous bound for κ(K) is given; the paper relies on numerical trials. The stated gate complexity for the coupled system is therefore conditional on this empirical observation.

pith-pipeline@v1.3.0-alltime-deepseek · 23977 in / 17506 out tokens · 180923 ms · 2026-08-01T04:52:47.109685+00:00 · methodology

0 comments
read the original abstract

The biharmonic equation is a prototypical fourth-order partial differential equation whose high-dimensional discretization suffers from rapidly growing degrees of freedom and severe ill-conditioning. We develop QSVT--VTAA quantum linear-system algorithms by constructing explicit block-encodings tailored to periodic, simply supported, and Dirichlet--Neumann boundary conditions. For periodic and simply supported problems, Fourier and sine-transform diagonalizations yield augmented Poisson systems with the condition-number scaling of a second-order operator. For Dirichlet--Neumann problems, we introduce a second-order boundary-corrected finite-difference discretization, establish mesh-independent stability, and construct an explicit block-encoding of the resulting nonsymmetric matrix. We also formulate a coupled-Laplace system with additional boundary unknowns and characterize its complexity in terms of the condition number of the complete augmented matrix. The analysis covers discretization error, block-encoding normalization, gate complexity, and solution extraction under an amplitude-input and quantum-state-output model. Numerical experiments validate the proposed discretizations and the corresponding linear solves.

Figures

Figures reproduced from arXiv: 2607.22396 by Chuwen Ma, Zihao Tang.

Figure 1
Figure 1. Figure 1: LCU construction with PREP, SELECT, and PREP† . superpositions over spatial directions, and their preparation cost is included in the stated block￾encoding costs. We will repeatedly use the following standard arithmetic rules for block-encodings [17, 23]. Lemma 2.2 (Arithmetic of block-encodings). Suppose UA is an (α, a, εA)-block-encoding of A with cost TA, and UB is a (β, b, εB)-block￾encoding of B with … view at source ↗
Figure 2
Figure 2. Figure 2: Basic QSVT circuit for applying a polynomial [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Boundary reflections R0 and RN−1 used in the shift-based block-encoding of the one-dimensional Dirichlet Laplacian. Lemma 2.5 (Discrete Laplacian block-encod￾ing). The one-dimensional Dirichlet Laplacian L (1) h admits an exact block-encoding with normal￾ization α1 = O(h −2 ) and gate cost O(n), up to the cost of multi￾controlled gates. Moreover, the d-dimensional tensor-product Laplacian L (d) h = X d ℓ=1… view at source ↗
Figure 4
Figure 4. Figure 4: Circuit structure for the periodic biharmonic [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Circuit structure for the simply supported bi [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Explicit quantum circuit for the discrete sine [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Block-encoding of Db. n V (ω) = P(2n−1ω) P(2n−2ω) P(ω) . . [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Block-encoding of V (ω). Lemma 4.1 (Block-encoding of the simply sup￾ported augmented matrix). The matrix DN in (22) admits an exact block-encoding with normal￾ization αDN = 4d(N + 1)2 + 1. Under a direct implementation of the direc￾tional SELECT operation, one use of this block￾encoding has gate cost TDN = O(d log N). Moreover, its normalization overhead satisfies βDN := αDN ∥DN ∥ = O(1). Proof. According… view at source ↗
Figure 9
Figure 9. Figure 9: Optimal block-encoding of matrix M The integer scaling corresponds to multiplying the row (11, −5, 5/3, −1/4) by 12; the normaliza￾tion constant and the global factor are absorbed into the block-encoding normalization αM = ∥M∥. By symmetry, (X ⊗ X)|uM⟩ encodes the last nonzero boundary row. Define |ϕ0⟩ = [PITH_FULL_IMAGE:figures/full_fig_p016_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Explicit constant-size circuit for preparing the [PITH_FULL_IMAGE:figures/full_fig_p017_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Block-encoding of matrix T Lemma 5.4 (Block-encoding of the direct dis￾cretization matrix). The matrix Ae in (31) admits an exact block-encoding with normalization αAe = 16d 2 (N + 1)4 + √ 21433 12 d(N + 1)4 . Under a direct implementation of the direc￾tional SELECT operation, one use of this block￾encoding has gate cost TAe = O(d log N). Moreover, its normalization overhead satisfies βAe := αAe [PITH_FU… view at source ↗
Figure 12
Figure 12. Figure 12: UG: block-encoding of the matrix G. Regarding matrix Aapm, it shares an analogous structure with AIB; hence, the methodology for constructing the block-encoding of AIB is directly applicable to Aapm. It should be noted that, un￾like AIB, Aapm requires padding with a bottom zero-block to be transformed into a square ma￾trix. We omit further details here for the sake of brevity. Once the block components ar… view at source ↗
Figure 13
Figure 13. Figure 13: Periodic problem (39). Comparison between the exact solution, the classical spectral discretization, and the QSVT circuit simulation. 7.2 Simply supported boundary conditions We next solve ( u (4)(x) = π 4 [PITH_FULL_IMAGE:figures/full_fig_p021_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Simply supported problem (41). The quan￾tum simulation and the classical finite-difference solution have comparable errors at the same grid resolution. 7.3 Direct Dirichlet–Neumann discretization To check the convergence of the direct discretiza￾tion, we consider the three-dimensional test prob￾lem with exact solution u(x, y, z) = x(1 − x) sin(πy) sin(πz). (43) The numerical results are summarized in [PI… view at source ↗
Figure 15
Figure 15. Figure 15: Convergence order for the coupled Laplace [PITH_FULL_IMAGE:figures/full_fig_p022_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Condition-number growth for the coupled Laplace system. 22 [PITH_FULL_IMAGE:figures/full_fig_p022_16.png] view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

35 extracted references · 7 canonical work pages

  1. [1]

    V. V. Meleshko. Selected topics in the history of the two-dimensional biharmonic problem.Applied Mechanics Reviews56, 33– 85 (2003)

    2003

  2. [2]

    Martinod, A

    T. Martinod, A. Fierro, C. Posada, and J. J. Sanchez. On the biharmonic equation. Cuadernos de Ingeniería Matemática1(4) (2024)

    2024

  3. [3]

    Ziegler.Mechanics of Solids and Fluids

    F. Ziegler.Mechanics of Solids and Fluids. Springer, New York (1995). doi:10.1007/978- 1-4612-0805-1

    doi:10.1007/978- 1995

  4. [4]

    Eschenauer, N

    H. Eschenauer, N. Olhoff, and W. Schnell. Applied Structural Mechanics: Fundamen- tals of Elasticity, Load-Bearing Structures, Structural Optimization. Springer, Berlin (1997). doi:10.1007/978-3-642-59205-8

    work page doi:10.1007/978-3-642-59205-8 1997

  5. [5]

    J. D. Anderson, Jr.Computational Fluid Dynamics: The Basics with Applications. McGraw–Hill, New York (1995)

    1995

  6. [6]

    Crane, U

    K. Crane, U. Pinkall, and P. Schröder. Ro- bust fairing via conformal curvature flow. ACM Transactions on Graphics32, 61:1– 61:10 (2013). doi:10.1145/2461912.2461986

    arXiv 2013

  7. [7]

    W. H. Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery.Numerical Recipes: The Art of Scientific Computing, 3rd ed. Cambridge University Press, Cam- bridge (2007)

    2007

  8. [8]

    J. H. Argyris, I. Fried, and D. W. Scharpf. The TUBA family of plate elements for the matrix displacement method.Aeronautical Journal72, 701–709 (1968)

    1968

  9. [9]

    K. Bell. A refined triangular plate bending finite element.International Journal for Nu- merical Methods in Engineering1, 101–122 (1969)

    1969

  10. [10]

    F. K. Bogner, R. L. Fox, and L. A. Schmit. The generation of interelement-compatible stiffness and mass matrices by the use of in- terpolation formulas. InProceedings of the Conference on Matrix Methods in Structural Mechanics, pp. 397–444, Air Force Institute of Technology, Wright–Patterson Air Force Base, Ohio (1965)

    1965

  11. [11]

    P. G. Ciarlet and P.-A. Raviart. A mixed finite element method for the biharmonic equation. In C. de Boor, ed.,Mathematical Aspects of Finite Elements in Partial Dif- ferential Equations, pp. 125–143. Academic Press, New York (1974). 23

    1974

  12. [12]

    L. W. Ehrlich. Solving the biharmonic equa- tion as coupled finite difference equations. SIAM Journal on Numerical Analysis8, 278–287 (1971)

    1971

  13. [13]

    P. E. Bjørstad. Fast numerical solution of the biharmonic Dirichlet problem on rectan- gles.SIAM Journal on Numerical Analysis 20, 59–71 (1983). doi:10.1137/0720004

    work page doi:10.1137/0720004 1983

  14. [14]

    J. W. Stephenson. Single cell discretiza- tions of order two and four for bihar- monic problems.Journal of Computational Physics55, 65–80 (1984). doi:10.1016/0021- 9991(84)90015-9

    doi:10.1016/0021- 1984

  15. [15]

    A. W. Harrow, A. Hassidim, and S. Lloyd. Quantum algorithm for lin- ear systems of equations.Physical Review Letters103, 150502 (2009). doi:10.1103/PhysRevLett.103.150502

    doi:10.1103/physrevlett.103.150502 2009

  16. [16]

    A. M. Childs, R. Kothari, and R. D. Somma. Quantum algorithm for systems of linear equations with exponentially im- proveddependenceonprecision.SIAM Jour- nal on Computing46, 1920–1950 (2017). doi:10.1137/16M1087072

    doi:10.1137/16m1087072 1920

  17. [17]

    Gilyén, Y

    A. Gilyén, Y. Su, G. H. Low, and N. Wiebe. Quantum singular value transformation and beyond: Exponential speedups for quantum algorithms. InProceedings of the 51st An- nual ACM SIGACT Symposium on Theory of Computing (STOC 2019), pp. 193–204 (2019). doi:10.1145/3313276.3316366

    arXiv 2019

  18. [18]

    Jin and N

    S. Jin and N. Liu. Quantum simulation of discrete linear dynamical systems and sim- ple iterative methods in linear algebra via Schrödingerisation.Proceedings of the Royal Society A: Mathematical, Physical and En- gineering Sciences480, 20230370 (2024). doi:10.1098/rspa.2023.0370

    arXiv 2024

  19. [19]

    S. Jin, N. Liu, C. Ma, and Y. Yu. Quan- tum preconditioning method for finite dif- ference discretizations of the Poisson equa- tion via Schrödingerization.arXiv preprint arXiv:2505.06866v2(2026)

    Pith/arXiv arXiv 2026

  20. [20]

    Chakraborty, A

    S. Chakraborty, A. Morolia, and A. Peduri. Quantum regularized least squares.Quan- tum7, 988 (2023). doi:10.22331/q-2023-04- 27-988

    work page doi:10.22331/q-2023-04- 2023

  21. [21]

    Kharazi, A

    T. Kharazi, A. M. Alkadri, J.-P. Liu, K. K. Mandadapu, and K. B. Whaley. Explicit block encodings of boundary value problems for many-body elliptic operators.Quantum 9, 1764 (2025). doi:10.22331/q-2025-06-04- 1764

    work page doi:10.22331/q-2025-06-04- 2025

  22. [22]

    G. H. Low and I. L. Chuang. Opti- mal Hamiltonian simulation by quan- tum signal processing.Physical Re- view Letters118, 010501 (2017). doi:10.1103/PhysRevLett.118.010501

    doi:10.1103/physrevlett.118.010501 2017

  23. [23]

    L. Lin. Lecture notes on quantum algorithms for scientific computation.arXiv preprint arXiv:2201.08309(2022). arXiv:2201.08309

    Pith/arXiv arXiv 2022

  24. [24]

    Möttönen, J

    M. Möttönen, J. J. Vartiainen, V. Bergholm, and M. M. Salomaa. Quantum circuits for general multiqubit gates.Physi- cal Review Letters93, 130502 (2004). doi:10.1103/PhysRevLett.93.130502

    doi:10.1103/physrevlett.93.130502 2004

  25. [25]

    L. Lin. Mathematical and numerical analysis of quantum signal processing.arXiv preprint arXiv:2510.00443(2025). arXiv:2510.00443

    arXiv 2025

  26. [26]

    C. Gidney. Halving the cost of quan- tum addition.Quantum2, 74 (2018). doi:10.22331/q-2018-06-18-74

    doi:10.22331/q-2018-06-18-74 2018

  27. [27]

    Novikau and I

    I. Novikau and I. Joseph. Estimating QSVT angles for matrix inversion with large con- dition numbers.Journal of Computational Physics525, 113767 (2025)

    2025

  28. [28]

    Klappenecker and M

    A. Klappenecker and M. Rötteler. Discrete cosine transforms on quantum computers. InProceedings of the Second International Symposium on Image and Signal Processing and Analysis, pp. 464–468 (2001)

    2001

  29. [29]

    A. A. Zecchi, C. Sanavio, L. Cappelli, S. Perotto, A. Roggero, and S. Succi. Block encoding of sparse matrices with a pe- riodic diagonal structure.arXiv preprint arXiv:2602.10589(2026). arXiv:2602.10589

    arXiv 2026

  30. [30]

    J. H. Bramble. A second order finite differ- ence analog of the first biharmonic bound- ary value problem.Numerische Mathematik 9, 236–249 (1966). doi:10.1007/BF02162087

    work page doi:10.1007/bf02162087 1966

  31. [31]

    Saad.Iterative Methods for Sparse Linear Systems, 2nd ed

    Y. Saad.Iterative Methods for Sparse Linear Systems, 2nd ed. SIAM, Philadelphia (2003). doi:10.1137/1.9780898718003

    doi:10.1137/1.9780898718003 2003

  32. [32]

    G. H. Golub and C. F. Van Loan.Matrix Computations, 4th ed. Johns Hopkins Uni- versity Press, Baltimore (2013)

    2013

  33. [33]

    Heinrichs

    W. Heinrichs. A stabilized treatment of the biharmonic operator with spectral meth- ods.SIAM Journal on Scientific and Sta- tistical Computing12, 1162–1172 (1991). doi:10.1137/0912061

    work page doi:10.1137/0912061 1991

  34. [34]

    Bialecki and A

    B. Bialecki and A. Karageorghis. A Legen- 24 dre spectral Galerkin method for the bihar- monic Dirichlet problem.SIAM Journal on Scientific Computing22, 1549–1569 (2000). doi:10.1137/S1064827598342407

    work page doi:10.1137/s1064827598342407 2000

  35. [35]

    UnitaryLab 2.0.https:// test-studio.UnitaryLab.com/

    UnitaryLab. UnitaryLab 2.0.https:// test-studio.UnitaryLab.com/. Accessed May 21, 2026. 25

    2026