Pith. sign in

REVIEW 3 major objections 4 minor 79 references

This paper claims that reverse-mode gradients for sparse Pauli dynamics can be computed with one backward pass, eliminating the n_param-fold memory overhead of cached automatic differentiation while preserving runtime.

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 ↗

A backward-propagation algorithm computes gradients for sparse Pauli dynamics in O(1) passes and O(N_P) memory, with gradient accuracy empirically comparable to the simulation's own energy accuracy.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection Solid exact reverse-mode gradient method for sparse Pauli dynamics, but the finite-truncation gradient accuracy claim is not supported by the empirical evidence provided. the 3 major comments →

arxiv 2607.15184 v1 pith:3PMMKD3F submitted 2026-07-16 quant-ph cond-mat.str-el

Backpropagating Pauli Propagation

Sheng-Hsuan Lin , Etienne Granet , Kevin H\'emery , Henrik Dreyer This is my paper
classification quant-ph cond-mat.str-el
keywords sparse Pauli dynamicsPauli propagationbackpropagationgradient computationquantum circuit optimizationreverse-mode automatic differentiationoperator stabilizer Rényi entropystate preparation
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

The paper develops a backpropagation algorithm for computing parameter gradients in quantum circuits simulated by sparse Pauli dynamics (SPD). Its central claim is that reverse-mode differentiation can be run without storing the forward trajectory: because the circuit is unitary, the intermediate operators are recomputed during a backward propagation sweep. This reduces gradient memory from O(N_P n_param) to O(N_P) with the same asymptotic runtime as one SPD evaluation, and gives approximate gradients whose accuracy is of the same order as the SPD estimate of the observable. The authors demonstrate the method by optimizing low-energy state-preparation circuits for Ising and Heisenberg models in one, two, and three dimensions, and by compressing two-dimensional time-evolution circuits.

Core claim

The paper's central claim is that reverse-mode gradient evaluation for sparse Pauli dynamics can be implemented without storing the full forward trajectory of intermediate operators. Because each Pauli rotation acts reversibly in the Heisenberg picture, the observable can be propagated backward through the Hermitian conjugate circuit, and the parameter gradients are accumulated at each gate from local cross terms of the current coefficients and coefficient derivatives. The result is a gradient computation that costs one extra sweep, uses memory O(N_P) instead of O(N_P n_param), and—under finite truncation—produces gradients whose error tracks the SPD expectation-value error rather than being

What carries the argument

The central object is the Pauli-rotation update rule of SPD: every Pauli string either commutes with the generator sigma and keeps its coefficient, or pairs with its conjugate string Q proportional to sigma P, and the two coefficients transform under the 2x2 rotation U(theta) = [[cos theta, sin theta], [-sin theta, cos theta]]. This rotation is reversible, so the backward pass applies its transpose and recomputes the preceding sparse operator on the fly. The gradient formula at each gate sums over anti-commuting pairs (P, Q) of (partial L / partial a_P) a_Q - (partial L / partial a_Q) a_P, which needs only the current coefficients and derivatives and therefore requires storing only one spars

Load-bearing premise

The load-bearing premise is that when truncating the backward pass by |a_P| < delta, a small coefficient in the forward observable generally has a small derivative with respect to each parameter; if a parameter is fine-tuned so a barely-discarded coefficient changes rapidly, the gradient is biased even when the expectation value is accurate.

What would settle it

Take a fixed 2D TFI HVA circuit from the paper's benchmark, set threshold delta = 1e-3, and compute the gradient both by the backpropagation algorithm and by exact zero-threshold SPD at the same parameters. If the gradient difference exceeds the observable error estimate, or if enumerating discarded conjugate pairs near the threshold reveals any string with |a_P| < delta but |partial theta a_P| large enough to matter, the correlation assumption fails.

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

If this is right

  • Gradient-based optimization of variational circuits becomes feasible at the memory limit of SPD, since the gradient computation costs one extra forward-like sweep and stores only one sparse operator and its derivatives.
  • Compared with finite-difference gradients, the method is O(n_param) cheaper in function evaluations; compared with standard reverse-mode automatic differentiation, it is O(n_param) cheaper in memory at the same asymptotic runtime.
  • When the coefficient threshold delta is lowered, both the expectation value and the gradient improve systematically, and the gradient error follows the observable error to the same order.
  • The same framework lets the operator stabilizer Renyi entropy be monitored and regularized during optimization, giving a concrete tradeoff between energy accuracy and operator non-stabilizerness.

Where Pith is reading between the lines

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

  • If the coefficient-derivative correlation holds broadly, SPD with backpropagation could replace stochastic gradient methods as the default classical optimizer for variational circuits, since it needs only one extra pass and no step-size tuning.
  • The paper's truncation heuristic suggests a testable improvement: keep a Pauli string when a weighted combination of |a_P| and |partial L / partial a_P| exceeds the threshold, which would convert the fine-tuning caveat into a controllable bias.
  • The OSE-regularization experiments imply an optimization strategy beyond the paper's demonstrations: one could scan the regularization strength to map how much operator non-stabilizerness is strictly necessary for a target energy, yielding a practical heuristic for low-magic circuit design.
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

3 major / 4 minor

Summary. The paper presents a reverse-mode backpropagation algorithm for evaluating parameter gradients in sparse Pauli dynamics (SPD). Instead of caching all intermediate sparse Pauli operators during a forward sweep, it recomputes them in a backward sweep by exploiting circuit reversibility, reducing memory from O(N_P n_param) to O(N_P) while keeping O(1) runtime relative to a single SPD evaluation. The exact gradient formulas are derived by the chain rule from the Pauli rotation update, and the paper claims that, with finite truncation, the resulting approximate gradients have accuracy of the same order as the SPD estimate of the observable expectation value. The method is benchmarked on state preparation for 1D, 2D, and 3D transverse-field Ising models and the 3D Heisenberg model, and on compression of 2D time-evolution circuits.

Significance. If the central approximate-gradient claim is established, this is a practically useful contribution: memory is the dominant bottleneck in large SPD simulations, and an O(n_param) memory reduction without asymptotic runtime overhead would make gradient-based SPD optimization far more scalable. The exact-algorithm part is clean and well motivated, and the 1D free-fermion benchmark reproduces published MPS energies to all reported digits, giving strong evidence that the exact gradient machinery is correct when truncation is absent. The paper also ships code and gives a careful comparison with finite-difference and cached reverse-mode approaches. However, the paper's headline accuracy claim for the finite-truncation regime is not yet backed by a proof or a direct numerical test against exact gradients; this is the main weakness.

major comments (3)
  1. [Sec. II.B.2, Eq. (21)] The central approximate-gradient claim is not established. The backward truncation discards the tuple (P, a_P, ∂L/∂a_P) whenever |a_P| < δ, but Eq. (21) evaluates the gradient from cross terms (∂L/∂a_P)a_Q − (∂L/∂a_Q)a_P, and Eq. (7) gives ∂a_P/∂θ = ±a_Q. Thus a small a_P can coexist with a large a_Q and a large gradient contribution when the 2×2 rotation nearly cancels the coefficient. The paper explicitly concedes this possibility, and the response is only a fine-tuning intuition, not a bound or a systematic numerical test. Since the abstract claims gradient accuracy is of the same order as the SPD expectation-value error, this heuristic is load-bearing. I ask for either a rigorous bound under stated conditions or a direct comparison against exact gradients (δ=0) on small systems, reporting worst-case ratios of gradient error to observable error across random parameters and thresholds.
  2. [Sec. III.A.1, Fig. 3] The validation in Fig. 3 compares the proposed backpropagation gradient against a BP reference with threshold 10^-12. This demonstrates that the approximate gradient converges to a high-precision BP estimate, but it does not validate the claim that the approximate gradient error is comparable to the SPD expectation-value error, because the reference uses the same truncation heuristic. Systematic biases common to SPD truncation are therefore invisible. Fig. 2, which tracks operator infidelity rather than gradient error, is likewise indirect. Please add a benchmark where exact gradients are available (e.g., free-fermion circuits with no truncation, or exact operator propagation for small N) and report gradient error and expectation-value error on the same circuits.
  3. [Table IV and Sec. III.A.3] The empirical error estimates for the 3D TFI results are internally inconsistent with the actual deviations from the QMC reference. For layer 2 the estimated truncation error is 2×10^-4 while the deviation is 5.6×10^-3; for layer 3 the estimate is 0.379 while the deviation is 1.5×10^-4 (layer 1 shows a similar discrepancy: 5×10^-6 estimated vs 0.019 actual). The statement that the two-layer result is 'close' to the QMC value is therefore not supported, and the layer-3 estimate is not a useful diagnostic. Since Eq. (B16) is presented as the practical error estimator and used throughout the results to claim controlled optimization, this discrepancy should be addressed, or the claims should be restricted to cases where the estimator has been validated.
minor comments (4)
  1. [Sec. II.B.2] The notation for backward-propagated coefficients is confusing: (P̃_i, ã_{P̃_i}, ∂L/∂ã_{P_i}) mixes tildes on Pauli labels and coefficients. Please define a single convention, e.g., ã_i with an explicit Pauli label.
  2. [Fig. 3] The caption says markers denote truncation thresholds δ=10^-2, 10^-4, 10^-6, but it is unclear how δ enters the finite-difference and central-difference curves. Please clarify whether δ is the SPD truncation threshold used in those evaluations.
  3. [Appendix B] Eq. (B16) is called an 'empirical error estimate' but is not guaranteed to be an upper bound, as the appendix itself explains. To avoid overclaiming, consider calling it a 'diagnostic' or 'heuristic indicator' throughout the paper.
  4. [Sec. III.A.1] The comparison with Ref. [25] uses an SPSA-optimized symmetry-breaking ansatz while the free-fermion benchmark uses HVA. The paper notes this, but the caption of Fig. 4(b) could be clearer that the SPSA data use a different ansatz and depth.

Circularity Check

0 steps flagged

No circularity: gradient formula follows by chain rule from the explicit Pauli update; self-citations are contextual, and the truncation heuristic is a stated limitation, not a circular input.

full rationale

The central derivation is self-contained. The backward propagation rule in Eq. (20) is the adjoint (transpose) chain rule applied to the explicit 2x2 rotation appearing in Eq. (7), and the parameter gradient in Eq. (21) is obtained by differentiating that rotation directly; no fitted constants, no target quantities, and no external theorem are used as inputs to the derivation. The memory saving claim is an algorithmic consequence of recomputing intermediate operators instead of caching them, not an identity between input and output. The self-citations used in the paper (e.g., Refs. [33], [40], and the code repository [70]) provide context, benchmark data, or closed-form reference energies and are not load-bearing premises of the backpropagation method; the central validation comparisons in Sec. III and Appendix A use external tensor-network and quantum Monte Carlo references. The paper's own caveats about the truncation rule—'it is possible that the gradient contribution ... is not small even when |a_P| < δ' and 'intuitively expected that a small a_P correlates with a small derivative'—are explicitly acknowledged heuristics, and Appendix B labels its error estimate 'an empirical observable-error estimate, not as a rigorous bound.' These are genuine limitations and should be weighed as correctness risk, but they are not circular: they do not assume the conclusion, and no parameter is fitted to the quantity being predicted. Overall, no step in the derivation reduces to its own input by construction.

Axiom & Free-Parameter Ledger

3 free parameters · 5 axioms · 0 invented entities

The method introduces no free physical parameters or invented entities. The derivation is closed-form, and the only input-dependent choices are numerical truncation controls and regularization strength. The main load-bearing heuristic is the gradient-truncation rule and the empirical error estimate, both flagged by the authors as non-rigorous.

free parameters (3)
  • SPD truncation threshold δ = 10^-2, 10^-3, 10^-4 in main experiments; 10^-6 and 10^-12 in tests
    Control parameter for sparse Pauli truncation. It sets both the simulation accuracy and the gradient accuracy; no automatic selection rule is given.
  • Maximum number of retained Pauli strings = Not specified precisely in main text
    Mentioned as an additional truncation control alongside δ; limits memory and runtime.
  • OSE regularization strength λ_OSE = 0 to 0.1 across experiments
    User-selected hyperparameter in the regularization demonstration; not fitted to data.
axioms (5)
  • standard math The circuit consists of Pauli rotation gates U_σ(θ) = exp(-iθσ/2), and the Heisenberg update of a Pauli string under such a gate is exactly Eq. (6).
    Used throughout Sec. II.A to derive the 2×2 rotation update and the gradient formula.
  • domain assumption Quantum circuits are unitary and hence reversible, so intermediate operators can be reconstructed exactly in the backward pass when no truncation is applied.
    Central to the memory-saving recomputation scheme in Sec. II.B.
  • domain assumption The cost function depends only on the truncated final SPO coefficients and is differentiable with respect to those coefficients.
    Assumed in Sec. II.B.2 to define the adjoint initialization ∂L/∂a_P^{(t)}.
  • ad hoc to paper Discarding backward derivative terms where the operator coefficient satisfies |a_P| < δ introduces negligible gradient error.
    The finite-truncation gradient accuracy rests on this heuristic, justified only by an intuitive fine-tuning argument in Sec. II.B.2.
  • ad hoc to paper The empirical SPD error estimate, root-sum-square accumulation of discarded ℓ2 weights (Eq. B16), is a usable proxy for observable error.
    Appendix B explicitly states this is not a rigorous bound and depends on unverified assumptions about alignment of truncation residuals.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Backpropagating Pauli Propagation." pith.science (2026). https://pith.science/paper/3PMMKD3F

@misc{pith2026260715184,
  author       = {Pith},
  title        = {Pith review of: Backpropagating Pauli Propagation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3PMMKD3F}},
  note         = {Machine review of arXiv:2607.15184}
}
Share X Bluesky LinkedIn Reddit HN
abstract

We develop a backpropagation algorithm for evaluating parameter gradients in quantum circuits using Pauli propagation simulation. The method has computational complexity comparable to that of standard sparse Pauli simulation techniques, while producing gradients whose accuracy is of the same order as the corresponding observable expectation values. By exploiting the reversibility of quantum circuits, the algorithm reduces the memory cost by a factor of $\mathcal{O}(n_\text{param})$ compared with conventional reverse-mode automatic differentiation, where $n_\text{param}$ denotes the number of parameters in the circuit. Compared with finite difference methods, the algorithm is $\mathcal{O}(n_\text{param})$ more efficient in function evaluations. These features enable efficient and accurate classical optimization of quantum circuits for applications such as state preparation and time-evolution compression, while also allowing operator-complexity measures such as the operator stabilizer R\'enyi entropy to be monitored and regularized during optimization. We demonstrate the method by optimizing low-energy state-preparation circuits for transverse-field Ising models in one, two, and three dimensions and for the three-dimensional Heisenberg model, and by compressing two-dimensional time-evolution circuits.

Figures

Figures reproduced from arXiv: 2607.15184 by Etienne Granet, Henrik Dreyer, Kevin H\'emery, Sheng-Hsuan Lin.

Figure 1
Figure 1. Figure 1: FIG. 1. An illustration of the forward propagation and backward propagation workflow of the SPD optimization. The workflow [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Error (a) and the number of strings (b) for forward [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. We compare the [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. (a) Relative energy error of the 1D TFI [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Relative energy error of the 2D TFI [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Time evolution circuit compression with 2D TFI [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Relative energy error and operator stabilizer entropy [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. Numerical test of the empirical SPD error estimate [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9. Average throughput for the two-dimensional global [PITH_FULL_IMAGE:figures/full_fig_p018_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10. Energy-density optimization results for the 1D TFI model using a 10-layer HVA circuit. For 10 fixed random [PITH_FULL_IMAGE:figures/full_fig_p019_10.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

79 extracted references · 10 linked inside Pith

  1. [1]

    The model has a quantum critical point atg c = 1

    Transverse-field Ising model in 1D In 1D, the TFI model can be mapped to a free fermion model through Jordan-Wigner transformation and is ex- actly solvable. The model has a quantum critical point atg c = 1. The ground states belong to aZ 2 symmetry broken ferromagnetic phase forg <1 and belong to a paramagnetic phase forg >1. As an exactly solvable model...

  2. [2]

    Given an application and problem of interest, the cost functionLquantifies the “error” of the solution associ- ated with the parametersθ

    Gradient Having defined both forward and backward propaga- tion, we are now in a position to show how to obtain the partial derivatives of the cost function with respect to the circuit parameters, i.e., the gradient. Given an application and problem of interest, the cost functionLquantifies the “error” of the solution associ- ated with the parametersθ. Th...

  3. [3]

    These include the finite- difference methods and the auto-differentiation methods

    Comparison with other differentiation schemes The proposed backpropagation algorithm should be compared with several alternatives for evaluating gra- dients in an SPD simulation. These include the finite- difference methods and the auto-differentiation methods. 7 For simplicity, in the following complexity comparison we assume that every parameterized gat...

  4. [4]

    We find the low-energy state prepa- ration circuit for 2D TFI withg= 3.1 near the crit- ical point by minimizing the energy expectation value

    Transverse-field Ising model in 2D After benchmarking with exact simulation in 1D, we turn to the 2D simulation where the SPD includes fi- nite truncation. We find the low-energy state prepa- ration circuit for 2D TFI withg= 3.1 near the crit- ical point by minimizing the energy expectation value. Here we consider the Hamiltonian variational ansatz as in ...

  5. [5]

    The purpose of these calcula- tions is not to establish state-of-the-art variational en- ergies, but to show that the method can easily be gen- eralized to different geometries

    Three-dimensional models Finally, we apply the same SPD backpropagation workflow to 3D lattices. The purpose of these calcula- tions is not to establish state-of-the-art variational en- ergies, but to show that the method can easily be gen- eralized to different geometries. For the 3D TFI model atg= 5.2, close to the critical point, a two-layer HV A circu...

  6. [6]

    The setup is as follows

    Transverse-field Ising model in 2D Below, we show a prototype test on the variational time evolution compression task. The setup is as follows. We consider a square lattice of size (10,10) with PBC and the target continuous time evolution unitaryV=e −i ˆHT , where ˆHis the 2D TFI withg= 3.1 and total time T= 0.3. We can estimate the light cone size by che...

  7. [7]

    The setup is as follows

    Regularization on 1D TFI circuits We demonstrate the effect of OSE regularization by revisiting the task of finding low-energy state prepara- tion circuit for the 1D TFI model forg= 1.1. The setup is as follows. We optimize the generic symmetry- breaking ansatz of Eq. 27 with the cost functionL total = E+λ OSE M[α=1]. We consider five-layer circuits and o...

  8. [8]

    forward pass

    Future work More broadly, the OSE perspective raises several open questions. We have demonstrated above using OSE regu- larization to probe the tradeoff between energy accuracy and operator non-stabilizerness. However, this should be distinguished from the more ambitious task of design- ing low-T-gate state-preparation algorithms. Converting this tradeoff...

  9. [9]

    J. C. Bridgeman and C. T. Chubb, Hand-waving and in- terpretive dance: an introductory course on tensor net- works, Journal of physics A: Mathematical and theoreti- cal50, 223001 (2017)

    2017

  10. [10]

    J. I. Cirac, D. Perez-Garcia, N. Schuch, and F. Ver- straete, Matrix product states and projected entangled pair states: Concepts, symmetries, theorems, Reviews of Modern Physics93, 045003 (2021)

    2021

  11. [11]

    Aaronson and D

    S. Aaronson and D. Gottesman, Improved simulation of stabilizer circuits, Physical Review A—Atomic, Molecu- lar, and Optical Physics70, 052328 (2004)

    2004

  12. [12]

    Bravyi, D

    S. Bravyi, D. Browne, P. Calpin, E. Campbell, D. Gosset, and M. Howard, Simulation of quantum circuits by low- rank stabilizer decompositions, Quantum3, 181 (2019)

    2019

  13. [13]

    Surace and L

    J. Surace and L. Tagliacozzo, Fermionic gaussian states: an introduction to numerical approaches, SciPost Physics Lecture Notes , 054 (2022)

    2022

  14. [14]

    Anand, K

    S. Anand, K. Temme, A. Kandala, and M. Zale- tel, Classical benchmarking of zero noise extrapola- tion beyond the exactly-verifiable regime, arXiv preprint arXiv:2306.17839 (2023)

    Pith/arXiv arXiv 2023

  15. [15]

    Beguˇ si´ c, J

    T. Beguˇ si´ c, J. Gray, and G. K.-L. Chan, Fast and con- verged classical simulations of evidence for the utility of quantum computing before fault tolerance, Science Ad- vances10, eadk4321 (2024)

    2024

  16. [16]

    Prosen and I

    T. Prosen and I. Piˇ zorn, Operator space entanglement entropy in a transverse ising chain, Physical Review A—Atomic, Molecular, and Optical Physics76, 032316 (2007)

    2007

  17. [17]

    Beguˇ si´ c, K

    T. Beguˇ si´ c, K. Hejazi, and G. K. Chan, Simulating quan- tum circuit expectation values by clifford perturbation theory, The Journal of Chemical Physics162(2025)

    2025

  18. [18]

    Dowling, P

    N. Dowling, P. Kos, and X. Turkeshi, Magic resources of the heisenberg picture, Physical Review Letters135, 050401 (2025)

    2025

  19. [19]

    Miller, J

    A. Miller, J. Favre, Z. Holmes, ¨O. Salehi, R. Chakraborty, A. Nyk¨ anen, Z. Zimbor´ as, A. Glos, and G. Garc ´ ıa-P´ erez, Simulation of fermionic circuits using majorana propaga- tion, arXiv preprint arXiv:2503.18939 (2025)

    arXiv 2025

  20. [20]

    Debertolis, Natural super-orbitals representation of many-body operators, arXiv preprint arXiv:2507.10690 (2025)

    M. Debertolis, Natural super-orbitals representation of many-body operators, arXiv preprint arXiv:2507.10690 (2025)

    Pith/arXiv arXiv 2025

  21. [21]

    Dowling, K

    N. Dowling, K. Modi, and G. A. White, Bridging en- tanglement and magic resources within operator space, Physical Review Letters135, 160201 (2025)

    2025

  22. [22]

    Gao and L

    X. Gao and L. Duan, Efficient classical simula- tion of noisy quantum computation, arXiv preprint arXiv:1810.03176 (2018)

    Pith/arXiv arXiv 2018

  23. [23]

    P. Rall, D. Liang, J. Cook, and W. Kretschmer, Simu- lation of qubit quantum circuits via pauli propagation, Physical Review A99, 062337 (2019)

    2019

  24. [24]

    M. S. Rudolph, T. Jones, Y. Teng, A. Angrisani, and Z. Holmes, Pauli propagation: A computational frame- work for simulating quantum systems, arXiv preprint arXiv:2505.21606 (2025)

    Pith/arXiv arXiv 2025

  25. [25]

    Aharonov, X

    D. Aharonov, X. Gao, Z. Landau, Y. Liu, and U. Vazi- rani, A polynomial-time classical algorithm for noisy ran- dom circuit sampling, inProceedings of the 55th Annual ACM Symposium on Theory of Computing(2023) pp. 945–957

    2023

  26. [26]

    N. A. Nemkov, E. O. Kiktenko, and A. K. Fedorov, Fourier expansion in variational quantum algorithms, Physical Review A108, 032406 (2023)

    2023

  27. [27]

    Y. Shao, F. Wei, S. Cheng, and Z. Liu, Simulating noisy variational quantum algorithms: A polynomial approach, Physical Review Letters133, 120603 (2024)

    2024

  28. [28]

    Schuster, C

    T. Schuster, C. Yin, X. Gao, and N. Y. Yao, A polynomial-time classical algorithm for noisy quantum circuits, Physical Review X15, 041018 (2025)

    2025

  29. [29]

    Angrisani, A

    A. Angrisani, A. Schmidhuber, M. S. Rudolph, M. Cerezo, Z. Holmes, and H.-Y. Huang, Classically esti- mating observables of noiseless quantum circuits, Physi- cal review letters135, 170602 (2025)

    2025

  30. [30]

    Fontana, M

    E. Fontana, M. S. Rudolph, R. Duncan, I. Rungger, and C. C ˆ ırstoiu, Classical simulations of noisy varia- tional quantum circuits, npj Quantum Information11, 84 (2025)

    2025

  31. [31]

    Cirstoiu, A fourier analysis framework for approx- imate classical simulations of quantum circuits, arXiv preprint arXiv:2410.13856 (2024)

    C. Cirstoiu, A fourier analysis framework for approx- imate classical simulations of quantum circuits, arXiv preprint arXiv:2410.13856 (2024)

    Pith/arXiv arXiv 2024

  32. [32]

    Beguˇ si´ c and G

    T. Beguˇ si´ c and G. K.-L. Chan, Real-time operator evo- lution in two and three dimensions via sparse pauli dy- namics, PRX quantum6, 020302 (2025)

    2025

  33. [33]

    C.-J. Lin, H. Gharibyan, and V. P. Su, Utility-scale quan- tum state preparation: Classical training using pauli path simulation, Quantum10, 2014 (2026)

    2014

  34. [34]

    D’Anna, Y

    M. D’Anna, Y. Zhang, R. Wiersema, M. S. Rudolph, and J. Carrasquilla, Circuit compression for 2d quantum dynamics, arXiv preprint arXiv:2507.01883 (2025)

    arXiv 2025

  35. [35]

    M. S. Rudolph, E. Fontana, Z. Holmes, and L. Cincio, Classical surrogate simulation of quantum systems with lowesa, arXiv preprint arXiv:2308.09109 (2023)

    Pith/arXiv arXiv 2023

  36. [36]

    Lerch, R

    S. Lerch, R. Puig, M. S. Rudolph, A. Angrisani, T. Jones, M. Cerezo, S. Thanasilp, and Z. Holmes, Efficient quantum-enhanced classical simulation for patches of quantum landscapes, PRX Quantum7, 020359 (2026)

    2026

  37. [37]

    Monaco, J

    S. Monaco, J. Slim, F. Rehm, D. Kr¨ ucker, and K. Borras, Symbolic pauli propagation for gradient- enabled pre-training of quantum circuits, arXiv preprint arXiv:2512.16674 (2025)

    Pith/arXiv arXiv 2025

  38. [38]

    Li and S.-X

    Z.-L. Li and S.-X. Zhang, Dual role of low-weight pauli propagation: A flawed simulator but a powerful initial- izer for variational quantum algorithms, Physical Review Research8, 013266 (2026)

    2026

  39. [39]

    A. G. Baydin, B. A. Pearlmutter, A. A. Radul, and J. M. Siskind, Automatic differentiation in machine learning: a survey, Journal of machine learning research18, 1 (2018)

    2018

  40. [40]

    Luo, J.-G

    X.-Z. Luo, J.-G. Liu, P. Zhang, and L. Wang, Yao.jl: Extensible, efficient framework for quantum algorithm design, arXiv preprint arXiv:1912.10877 (2019)

    Pith/arXiv arXiv 1912

  41. [41]

    S.-H. Lin, R. Dilip, A. G. Green, A. Smith, and F. Poll- mann, Real-and imaginary-time evolution with com- pressed quantum circuits, PRX Quantum2, 010342 (2021)

    2021

  42. [42]

    Jobst, A

    B. Jobst, A. Smith, and F. Pollmann, Finite-depth scal- ing of infinite quantum circuits for quantum critical points, Physical Review Research4, 033118 (2022)

    2022

  43. [43]

    Gibbs and L

    J. Gibbs and L. Cincio, Deep circuit compression for quantum dynamics via tensor networks, Quantum9, 1789 (2025)

    2025

  44. [44]

    Here the sum is understood to include each conjugate Pauli pair once; equivalently, an implementation may 22 enumerate both orientations and divide the result by two

  45. [45]

    When parameters are shared, the gate-level derivatives are summed over all gates: the tangent dimension in forward mode is set by the number of independent pa- rameters, whereas the storage required by a conventional cached reverse-mode implementation is set by the num- ber of parameterized gates

  46. [46]

    While we utilize the spatial average to reduce the com- putational complexity, the method applies equally well to non-translational invariant system

  47. [47]

    Sokolov and J

    I. Sokolov and J. Dziarmaga, Bang-bang preparation of a quantum many-body ground state in a finite lattice: Op- timization of the algorithm with a tensor network, Phys- ical Review B111, 245144 (2025)

    2025

  48. [48]

    Dreyer, M

    H. Dreyer, M. Bejan, and E. Granet, Quantum comput- ing critical exponents, Physical Review A104, 062614 (2021)

    2021

  49. [49]

    Park, Efficient ground state preparation in varia- tional quantum eigensolver with symmetry-breaking lay- ers, APL Quantum1(2024)

    C.-Y. Park, Efficient ground state preparation in varia- tional quantum eigensolver with symmetry-breaking lay- ers, APL Quantum1(2024)

    2024

  50. [50]

    There are several other possible benchmarks. For exam- ple, closed-form expressions for the energy expectation value are available [40], and for the depths considered here the light cone is small enough for exact statevec- tor simulation. For simplicity, we compare directly with existing data from the literature

  51. [51]

    [25] considers a finite PBC system of sizeN= 100

    Ref. [25] considers a finite PBC system of sizeN= 100. The comparison is valid for two reasons. Firstly, the finite ground-state correlation length,ξ∼10.4 [39], renders the energy density of a finite size system indistinguish- able from the thermodynamic limit. Secondly, the circuit light cone remains smaller than the system size, ensuring that the energy...

  52. [52]

    Zhang and J

    Y. Zhang and J. Dziarmaga, Bang-bang preparation of quantum many-body ground states in two dimensions: Optimization of the algorithm with a two-dimensional tensor network, Physical Review B109, 235124 (2024)

    2024

  53. [53]

    [25] were obtained and eval- uated on a finite 10×10 PBC lattice, and their relative errors use the corresponding finite-size reference energy

    The SPD–SPSA data of Ref. [25] were obtained and eval- uated on a finite 10×10 PBC lattice, and their relative errors use the corresponding finite-size reference energy. The SPD-backpropagation and tensor-network data in- stead use the iPEPS thermodynamic-limit reference en- ergy

  54. [54]

    P. C. Vlaar and P. Corboz, Simulation of three- dimensional quantum systems with projected entangled- pair states, Physical Review B103, 205137 (2021)

    2021

  55. [55]

    Bauer, L

    B. Bauer, L. D. Carr, H. G. Evertz, A. Feiguin, J. Freire, S. Fuchs, L. Gamper, J. Gukelberger, E. Gull, S. Guertler, A. Hehn, R. Igarashi, S. V. Isakov, D. Koop, P. N. Ma, P. Mates, H. Matsuo, O. Parcol- let, G. Pawlowski, J. D. Picon, L. Pollet, E. Santos, V. W. Scarola, U. Schollw¨ ock, C. Silva, B. Surer, S. Todo, S. Trebst, M. Troyer, M. L. Wall, P. ...

    2011

  56. [56]

    Mansuroglu, T

    R. Mansuroglu, T. Eckstein, L. N¨ utzel, S. A. Wilkin- son, and M. J. Hartmann, Variational hamiltonian simu- lation for translational invariant systems via classical pre- processing, Quantum Science and Technology8, 025006 (2023)

    2023

  57. [57]

    Mizuta and T

    K. Mizuta and T. Kuwahara, Trotterization is substan- tially efficient for low-energy states, Phys. Rev. Lett.135, 130602 (2025)

    2025

  58. [58]

    Verstraete and J

    F. Verstraete and J. I. Cirac, Matrix product states represent ground states faithfully, Physical Review B—Condensed Matter and Materials Physics73, 094423 (2006)

    2006

  59. [59]

    Schuch, M

    N. Schuch, M. M. Wolf, F. Verstraete, and J. I. Cirac, En- tropy scaling and simulability by matrix product states, Physical review letters100, 030504 (2008)

    2008

  60. [60]

    Other possible choices include|θ| 2 or the distance to a Clifford point

  61. [61]

    P. S. Tarabunga, E. Tirrito, T. Chanda, and M. Dal- monte, Many-body magic via pauli-markov chains—from criticality to gauge theories, PRX Quantum4, 040317 (2023)

    2023

  62. [62]

    Odavi´ c, T

    J. Odavi´ c, T. Haug, G. Torre, A. Hamma, F. Franchini, and S. M. Giampaolo, Complexity of frustration: A new source of non-local non-stabilizerness, SciPost physics15, 131 (2023)

    2023

  63. [63]

    P. R. N. Falc˜ ao, P. S. Tarabunga, M. Frau, E. Tirrito, J. Zakrzewski, and M. Dalmonte, Nonstabilizerness in u (1) lattice gauge theory, Physical Review B111, L081102 (2025)

    2025

  64. [64]

    Hoshino, M

    M. Hoshino, M. Oshikawa, and Y. Ashida, Stabilizer r´ enyi entropy and conformal field theory, Physical Re- view X16, 011037 (2026)

    2026

  65. [65]

    Timsina, Y.-M

    H. Timsina, Y.-M. Ding, E. Tirrito, P. S. Tarabunga, B.-B. Mao, M. Collura, Z. Yan, and M. Dalmonte, Robustness of magic in the quantum ising chain via quantum monte carlo tomography, arXiv preprint arXiv:2507.12902 (2025)

    arXiv 2025

  66. [66]

    J. W. Keeble, A. Lovato, and C. E. Robin, Neural quan- tum states in non-stabilizer regimes: Benchmarks with atomic nuclei, arXiv preprint arXiv:2603.28646 (2026)

    arXiv 2026

  67. [67]

    Turkeshi, E

    X. Turkeshi, E. Tirrito, and P. Sierant, Magic spread- ing in random quantum circuits, Nature Communications 16, 2575 (2025)

    2025

  68. [68]

    Turkeshi, A

    X. Turkeshi, A. Dymarsky, and P. Sierant, Pauli spec- trum and nonstabilizerness of typical quantum many- body states, Physical Review B111, 054301 (2025)

    2025

  69. [69]

    Tirrito, X

    E. Tirrito, X. Turkeshi, and P. Sierant, Anticoncentration and nonstabilizerness spreading under ergodic quantum dynamics, arXiv preprint arXiv:2412.10229 (2024)

    arXiv 2024

  70. [70]

    P. R. N. Falc˜ ao, P. Sierant, J. Zakrzewski, and E. Tir- rito, Nonstabilizerness dynamics in many-body localized systems, arXiv preprint arXiv:2503.07468 (2025)

    arXiv 2025

  71. [71]

    Aditya, A

    S. Aditya, A. Summer, P. Sierant, and X. Turkeshi, Mpemba effects in quantum complexity, arXiv preprint arXiv:2509.22176 (2025)

    arXiv 2025

  72. [72]

    Aditya, X

    S. Aditya, X. Turkeshi, and P. Sierant, Growth and spreading of quantum resources under random circuit dy- namics, arXiv preprint arXiv:2512.14827 (2025)

    arXiv 2025

  73. [73]

    Loizeau, J

    N. Loizeau, J. C. Peacock, and D. Sels, Quantum many- body simulations with paulistrings. jl, SciPost Physics Codebases , 054 (2025)

    2025

  74. [74]

    Broers, R.-Y

    L. Broers, R.-Y. Sun, and S. Yunoki, Scalable simulation of quantum many-body dynamics with or-represented quantum algebra, arXiv preprint arXiv:2506.13241 (2025)

    Pith/arXiv arXiv 2025

  75. [75]

    Bayraktar, A

    H. Bayraktar, A. Charara, D. Clark, S. Cohen, T. Costa, Y.-L. L. Fang, Y. Gao, J. Guan, J. Gunnels, A. Haidar, et al., cuquantum sdk: A high-performance library for accelerating quantum science, in2023 IEEE Interna- tional Conference on Quantum Computing and Engineer- ing (QCE), Vol. 1 (IEEE, 2023) pp. 1050–1061. 23

    2023

  76. [76]

    Watanabe, D

    R. Watanabe, D. Sels, and J. Tindall, Tensor network surrogate models for variational quantum computation, arXiv preprint arXiv:2604.20180 (2026)

    Pith/arXiv arXiv 2026

  77. [77]

    Miao and T

    Q. Miao and T. Barthel, Convergence and quantum ad- vantage of trotterized mera for strongly-correlated sys- tems, Quantum9, 1631 (2025)

    2025

  78. [78]

    Lin, Sparse pauli dynamics,https://github.com/ quantinuum/spd(2026)

    S.-H. Lin, Sparse pauli dynamics,https://github.com/ quantinuum/spd(2026)

    2026

  79. [79]

    M. C. Caro, H.-Y. Huang, N. Ezzell, J. Gibbs, A. T. Sornborger, L. Cincio, P. J. Coles, and Z. Holmes, Out- of-distribution generalization for learning quantum dy- namics, Nature Communications14, 3751 (2023)

    2023

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.