pith. sign in

arxiv: 2605.20331 · v1 · pith:2JZ57BD4new · submitted 2026-05-19 · 🪐 quant-ph

Bowtie VarQTE: A Resource-Efficient Quantum State Preparation Primitive

Marc Drudis , Alberto Baiardi , Mattia Chiurco , Francesco Tacchino , Stefan Woerner , Christa Zoufal This is my paper

Pith reviewed 2026-05-21 01:18 UTC · model grok-4.3

classification 🪐 quant-ph
keywords VarQTEquantum state preparationcausal light-conesvariational time evolutionresource efficiencyhybrid classical-quantum simulation2D quantum systems
0
0 comments X

The pith

Bowtie VarQTE prepares quantum states by classically simulating subcircuits inside causal light-cones while using quantum hardware only for the connected parts.

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

The paper presents bowtie VarQTE as a variational method to approximate state preparation through short-time real or imaginary evolution. It splits the circuit according to causal light-cones so that causally disconnected subcircuits run on classical simulators while only the relevant slices consume quantum resources. This removes the requirement for a complete classical representation of the target state that limits competing approaches such as approximate quantum compilation. Experiments on two-dimensional systems indicate that the method reaches similar fidelities to those alternatives yet lowers the number of quantum operations compared with standard sample-based Krylov methods.

Core claim

Bowtie VarQTE uses the existing causal light-cone structure of the underlying Hamiltonian to evaluate the gradient and quantum geometric tensor with a hybrid classical-quantum procedure, which permits exact McLachlan variational updates and thereby reduces overall quantum resource demands for preparing physically structured states.

What carries the argument

The bowtie circuit structure that isolates causally relevant subcircuits for classical simulation in gradient and quantum geometric tensor calculations.

Load-bearing premise

The target states must have causal light-cone structure strong enough that classical simulation of the disconnected subcircuits remains accurate.

What would settle it

A side-by-side run on the same 2D lattice where bowtie VarQTE consumes more quantum gates than a standard Trotter-compiled preparation circuit would refute the resource-reduction claim.

Figures

Figures reproduced from arXiv: 2605.20331 by Alberto Baiardi, Christa Zoufal, Francesco Tacchino, Marc Drudis, Mattia Chiurco, Stefan Woerner.

Figure 1
Figure 1. Figure 1: FIG. 1: Construction of a bowtie circuit for gradient [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: highlights how the infidelity between the pre￾pared and the target states behaves throughout the evo￾lution for: (a) an N = 35 qubit system with depth L = 4 for AQC as well as VarQRTE and L = 8 for Trotter, (b) N = 50 qubits with depth L = 3 for AQC as well as VarQRTE and L = 6 for Trotter. In both cases, AQC and [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3: Quantum circuit ansatz. The blocks of gates [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4: Energy estimates along the SKQD pipeline. [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5: Resources needed for the computation of [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6: Comparison of the QGT approximation error [PITH_FULL_IMAGE:figures/full_fig_p018_6.png] view at source ↗
read the original abstract

The preparation of quantum states is a fundamental requirement for many quantum algorithms. A native route to preparing physically structured states is based on short-time simulation of dynamical processes, such as real or imaginary time evolution. This work presents a resource-efficient framework for the approximation thereof with \textit{bowtie \ac{VarQTE}} which uses classical simulation where possible and quantum resources where necessary. We introduce a framework that leverages existing causal light-cones to minimize quantum resource requirements in the evaluation of gradient and quantum geometric tensor terms by utilizing classical simulation methods for causally relevant subcircuits. This in turn enables exact parameter updates according to McLachlan's variational principle and, thereby, improves numerical stability. We conduct a comparison with a state preparation method that is based on a tensor-network compiled Trotter algorithm: approximate quantum compilation (AQC). In recent work, this approach has shown impressive performance. However, its key-bottleneck is the necessity to have a classical (approximate) representation of the target state. Our numerical experiments indicate that bowtie VarQTE can achieve comparable fidelities without this requirement. We further illustrate how bowtie VarQTE can facilitate a state-preparation pipeline that combines the simulation of imaginary and real time evolution for a sample-based quantum algorithm. In fact, results on 2D systems show how bowtie VarQTE can reduce the quantum requirements compared to standard, sample-based Krylov diagonalization calculations. Our results indicate that VarQTE is a promising primitive for the preparation of physically structured quantum states that reduces requirements on quantum resources by leveraging existing structures and the associated possibility of enabling classical simulations.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 2 minor

Summary. The manuscript introduces bowtie VarQTE, a variational quantum time evolution framework for quantum state preparation. It exploits causal light-cones to classically simulate causally relevant subcircuits when evaluating gradients and the quantum geometric tensor, enabling exact McLachlan variational updates. Numerical experiments claim comparable fidelities to approximate quantum compilation (AQC) without requiring a classical representation of the target state, and demonstrate reduced quantum resource requirements relative to sample-based Krylov diagonalization for 2D systems by combining imaginary- and real-time evolution.

Significance. If the reported fidelities and resource reductions are robust, the method could serve as a practical hybrid primitive for preparing structured quantum states on near-term hardware by selectively offloading subcircuit evaluations to classical simulation. The avoidance of a full classical target-state representation distinguishes it from AQC and may broaden applicability to systems where such representations are costly.

major comments (2)
  1. [2D numerical results] Results section on 2D systems: the claim that bowtie VarQTE reduces quantum requirements compared to sample-based Krylov diagonalization is central to the resource-efficiency thesis, yet no explicit scaling of light-cone qubit count, classical simulation cost, or total gate depth versus system size or Trotter steps is provided. Without this, it remains unclear whether the observed fidelity parity arises from genuine net quantum savings or from a different variational parameterization once entanglement spreads.
  2. [Bowtie VarQTE framework] Method section describing the bowtie construction: the assertion that classical simulation of causally relevant subcircuits yields exact McLachlan updates for the full circuit rests on the light-cone remaining small enough for efficient classical contraction. In 2D imaginary-time evolution this assumption may break as the cone grows with both Trotter depth and ansatz range; the manuscript supplies no quantitative bound or counter-example demonstrating that the classical overhead stays sub-dominant.
minor comments (2)
  1. [Numerical experiments] Numerical experiments: fidelity and resource plots lack error bars, number of independent runs, and any statement on data exclusion or convergence criteria, making it difficult to assess the statistical significance of the reported parity with AQC and Krylov baselines.
  2. [Framework description] Notation: the definition of the bowtie circuit and the precise partitioning into classically simulable versus quantum-measured subcircuits is introduced without an accompanying diagram or explicit index notation for the light-cone boundaries, complicating reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their insightful comments and the opportunity to improve our manuscript. We address the major comments point-by-point below, providing clarifications and indicating revisions to be incorporated in the updated version.

read point-by-point responses

  1. Referee: Results section on 2D systems: the claim that bowtie VarQTE reduces quantum requirements compared to sample-based Krylov diagonalization is central to the resource-efficiency thesis, yet no explicit scaling of light-cone qubit count, classical simulation cost, or total gate depth versus system size or Trotter steps is provided. Without this, it remains unclear whether the observed fidelity parity arises from genuine net quantum savings or from a different variational parameterization once entanglement spreads.

    Authors: We acknowledge that explicit scaling analysis would better support the resource-efficiency claims. The numerical results in the manuscript demonstrate reduced quantum requirements for the specific 2D system sizes and evolution steps considered, where the light-cone sizes allow for classical simulation of subcircuits. To address this, we will revise the results section to include a discussion of the scaling behavior, including estimates of light-cone qubit counts as a function of system size and Trotter steps. For the local Hamiltonians and ansatz depths used, the light-cone remains bounded, leading to classical costs that do not offset the quantum savings. We will also clarify that the variational parameterization is consistent with the Krylov approach, with the savings coming from the hybrid evaluation rather than a different ansatz. This revision will be made. revision: yes

  2. Referee: Method section describing the bowtie construction: the assertion that classical simulation of causally relevant subcircuits yields exact McLachlan updates for the full circuit rests on the light-cone remaining small enough for efficient classical contraction. In 2D imaginary-time evolution this assumption may break as the cone grows with both Trotter depth and ansatz range; the manuscript supplies no quantitative bound or counter-example demonstrating that the classical overhead stays sub-dominant.

    Authors: The bowtie VarQTE framework is designed such that the causal light-cones are exploited precisely when they are small enough for classical simulation, which is the case in the early stages of imaginary-time evolution before significant entanglement spreads. We agree that a quantitative bound would strengthen the method section. In the revised manuscript, we will add a subsection providing a bound on the light-cone size based on the Lieb-Robinson bound or light-cone arguments for the 2D lattice, showing that for the Trotter steps and ansatz ranges in our experiments, the classical contraction cost remains polynomial and sub-dominant to the full quantum circuit cost. We will also discuss the regime where this holds and note that for deeper evolutions, hybrid fallback strategies can be employed. This will demonstrate that the assumption is valid for the presented results. revision: yes

Circularity Check

0 steps flagged

No circularity: bowtie VarQTE derivation is methodologically self-contained

full rationale

The paper introduces bowtie VarQTE as a new variational framework that exploits causal light-cones to enable classical simulation of relevant subcircuits when computing gradients and the quantum geometric tensor for McLachlan updates. Numerical comparisons to AQC (which requires a classical target-state representation) and to sample-based Krylov diagonalization are presented as external benchmarks, with reported 2D fidelity and resource outcomes arising from the proposed ansatz and simulation strategy rather than from any fitted parameter or self-referential definition. No equations or claims reduce the central resource-saving assertion to quantities defined by construction within the same work; the derivation chain therefore remains independent of its own outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only review yields limited visibility into parameters or assumptions; the central claims rest on the applicability of McLachlan's variational principle and the existence of usable causal light-cones in the target circuits.

axioms (1)
  • domain assumption McLachlan's variational principle can be applied to obtain exact parameter updates in the bowtie VarQTE setting
    Invoked to enable stable updates and improve numerical stability

pith-pipeline@v0.9.0 · 5845 in / 1298 out tokens · 35505 ms · 2026-05-21T01:18:33.246275+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

56 extracted references · 56 canonical work pages · 3 internal anchors

  1. [1]

    This capability, in turn, en- ables a systematic study of VarQTE at scale, based on causal locality:bowtie VarQTE

    and effectively isolate the relevant sub-systems such that the problem is decomposed into the independent simulation of smaller, tractable sub-circuits which can be analytically computed. This capability, in turn, en- ables a systematic study of VarQTE at scale, based on causal locality:bowtie VarQTE. The underlying frame- work allows the balanced use of ...

    work page

  2. [2]

    Mathematically, this involves approximating the exact evolution operatorU(t) =e −iHt by a variational ansatz, e.g., of the form given in Definition 2

    Approximate Quantum Compilation The core idea behind AQC is to simulate the unitary evolution using an Matrix Product State (MPS) which is 5 then compressed it into a parameterized circuit of fixed depth. Mathematically, this involves approximating the exact evolution operatorU(t) =e −iHt by a variational ansatz, e.g., of the form given in Definition 2. I...

    work page

  3. [3]

    Comparison We compare bowtie VarQRTE with a second-order Suzuki–Trotter baseline and an AQC-compiled version thereof on one-dimensional Heisenberg chains. The sim- ulated model is the nearest-neighbor Heisenberg Hamil- tonian on a 1D lattice, H1D = N−1X i=1 (XiXi+1 +Y iYi+1 +Z iZi+1),(17) and for both AQC and VarQRTE we use a Hamilto- nian variational ans...

    work page

  4. [4]

    This method leverages real-time evolution to gen- erate quantum states to be sampled from and construct a reduced subspace

    Sample-based Krylov Quantum Diagonalization SKQD combines elements from Sample Based Quan- tum Diagonalization [28] and Krylov subspace expansion [29]. This method leverages real-time evolution to gen- erate quantum states to be sampled from and construct a reduced subspace. Within this subspace, the Hamilto- nian is diagonalized to obtain an approximatio...

    work page

  5. [5]

    System Design & Setup Bowtie VarQTE provides a particularly natural front- end for SKQD. Imaginary VarQTE enables an initializa- tion strategy to prepare a system with a good ground- state overlap and a subsequent real-time VarQTE stage can help to generate a Krylov family{|ψ tk ⟩}. Moreover, if one wishes to probe real-time evolutions beyond those access...

    work page

  6. [6]

    Moreover, bowtie VarQRTE is subsequently used to generate a Krylov basis using shallow quantum cir- cuits

    Experiments We now present a proof-of-principle realization of the SKQD pipeline in which bowtie VarQITE is used to prepare an initial state that overlaps with the ground- state. Moreover, bowtie VarQRTE is subsequently used to generate a Krylov basis using shallow quantum cir- cuits. Samples collected from these circuits define a com- pact classical subs...

    work page

  7. [7]

    H. F. Trotter, Proceedings of the American Mathematical Society10, 545 (1959)

    work page 1959

  8. [8]

    Suzuki, Communications in Mathematical Physics51, 183 (1976)

    M. Suzuki, Communications in Mathematical Physics51, 183 (1976)

    work page 1976

  9. [9]

    Lloyd, Science273, 1073 (1996)

    S. Lloyd, Science273, 1073 (1996)

    work page 1996

  10. [10]

    Motta and et al., Nature Physics16, 205 (2020)

    M. Motta and et al., Nature Physics16, 205 (2020)

    work page 2020

  11. [11]

    Theory of variational quantum simulation,

    X. Yuan, S. Endo, Q. Zhao, Y. Li, and S. Benjamin, “Theory of variational quantum simulation,” (2019), arXiv:1812.08767 [quant-ph]

    work page arXiv 2019

  12. [12]

    McArdle, T

    S. McArdle, T. Jones, S. Endo, Y. Li, S. C. Benjamin, and X. Yuan, npj Quantum Information5, 75 (2019)

    work page 2019

  13. [13]

    McLachlan, Molecular Physics8(1964)

    A. McLachlan, Molecular Physics8(1964)

    work page 1964

  14. [14]

    Martinazzo and I

    R. Martinazzo and I. Burghardt, Phys. Rev. Lett.124, 150601 (2020)

    work page 2020

  15. [15]

    Zoufal, D

    C. Zoufal, D. Sutter, and S. Woerner, Physical Review Applied20(2023), 10.1103/physrevapplied.20.044059

    work page doi:10.1103/physrevapplied.20.044059 2023

  16. [16]

    Lightcone shading for classically accelerated quantum error mitigation,

    A. Eddins, M. C. Tran, and P. Rall, “Lightcone shading for classically accelerated quantum error mitigation,” (2024), arXiv:2409.04401 [quant-ph]

    work page arXiv 2024

  17. [17]

    Robertson, A

    N. Robertson, A. Akhriev, J. Vala, and S. Zhuk, ACM Transactions on Quantum Computing6, 1–15 (2025)

    work page 2025

  18. [18]

    Preparing 100-qubit symmetry-protected topological order on a digital quan- tum computer,

    G. Pennington, K. C. Smith, J. R. Garrison, L. P. Lin- doy, J. Crain, and B. Jaderberg, “Preparing 100-qubit symmetry-protected topological order on a digital quan- tum computer,” (2026), arXiv:2603.06325 [quant-ph]

    work page arXiv 2026

  19. [19]

    Kovyrshin, D

    A. Kovyrshin, D. Manawadu, E. Altamura, G. Penning- ton, and et al., Phys. Chem. Chem. Phys.28, 3035 (2026)

    work page 2026

  20. [20]

    D. E. Deutsch, Proceedings of the Royal Society of Lon- don. A. Mathematical and Physical Sciences425, 73 (1989), https://royalsocietypublishing.org/rspa/article- pdf/425/1868/73/67392/rspa.1989.0099.pdf

    work page arXiv 1989

  21. [21]

    Gacon, J

    J. Gacon, J. Nys, R. Rossi, S. Woerner, and G. Carleo, Phys. Rev. Res.6, 013143 (2024)

    work page 2024

  22. [22]

    Spiking neural network hypergraphs with spike frequency data,

    M. Drudis, “Bowtieqgt: Efficient quantum geometric ten- sor computation,”https://doi.org/10.5281/zenodo. 20210713(2026)

    work page doi:10.5281/zenodo 2026

  23. [23]

    Variational preparation of normal ma- trix product states on quantum computers,

    B. Jaderberg, G. Pennington, K. V. Marshall, L. W. An- derson, A. Agarwal, L. P. Lindoy, I. Rungger, S. Mensa, and J. Crain, “Variational preparation of normal ma- trix product states on quantum computers,” (2025), arXiv:2503.09683 [quant-ph]

    work page arXiv 2025

  24. [24]

    Sharma, S

    K. Sharma, S. Khatri, M. Cerezo, and P. J. Coles, New Journal of Physics22, 043006 (2020)

    work page 2020

  25. [25]

    Jones and S

    T. Jones and S. C. Benjamin, Quantum6, 628 (2022)

    work page 2022

  26. [26]

    Schollw¨ ock, Ann

    U. Schollw¨ ock, Ann. Phys.326, 96 (2011)

    work page 2011

  27. [27]

    Vidal, Phys

    G. Vidal, Phys. Rev. Lett.91, 147902 (2003)

    work page 2003

  28. [28]

    Sch¨ on, E

    C. Sch¨ on, E. Solano, F. Verstraete, J. I. Cirac, and M. M. Wolf, Phys. Rev. Lett.95, 110503 (2005)

    work page 2005

  29. [29]

    Qiskit addon: Approximate Quantum Com- pilation with Tensor Networks,

    J. R. Garrison, K. Marshall, I. Shehzad, K. J. Sung, and et al., “Qiskit addon: Approximate Quantum Com- pilation with Tensor Networks,”https://github.com/ Qiskit/qiskit-addon-aqc-tensor(2024)

    work page 2024

  30. [30]

    Quantum computing with Qiskit

    A. Javadi-Abhari and et al., “Quantum computing with Qiskit,” (2024), arXiv:2405.08810 [quant-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  31. [31]

    E. N. Epperly, L. Lin, and Y. Nakatsukasa, SIAM Jour- nal on Matrix Analysis and Applications43, 1263 (2022), https://doi.org/10.1137/21M145954X

    work page doi:10.1137/21m145954x 2022

  32. [33]

    Yoshioka, M

    N. Yoshioka, M. Amico, W. Kirby, P. Jurcevic, and et al., Nature Communications16, 5014 (2025)

    work page 2025

  33. [34]

    Robledo-Moreno, M

    J. Robledo-Moreno, M. Motta, H. Haas, A. Javadi- Abhari, and et al., Science Advances11(2025), 10.1126/sciadv.adu9991

    work page doi:10.1126/sciadv.adu9991 2025

  34. [35]

    A. N. Krylov, Izvestiya Akademii Nauk SSSR (1931), original work introducing what are now known as Krylov subspaces

    work page 1931

  35. [36]

    K¨ ohler, R

    F. K¨ ohler, R. Mukherjee, and P. Schmelcher, Phys. Rev. Res.5, 023135 (2023)

    work page 2023

  36. [37]

    T. Hu, Y. Gao, Y. Zhang, and et al., Quantum Informa- tion Processing (2024), 10.1007/s11128-024-04332-x

    work page doi:10.1007/s11128-024-04332-x 2024

  37. [38]

    Huang, X

    J. Huang, X. Qian, and M. Qin, Physical Review A112, 012425 (2025)

    work page 2025

  38. [39]

    L. E. Fischer, M. Leahy, A. Eddins, N. Keenan, and et al., Nature Physics22, 302 (2026)

    work page 2026

  39. [40]

    Qiskit addon: sample-based quan- tum diagonalization,

    A. A. Saki, S. Barison, B. Fuller, J. R. Garrison, and et al., “Qiskit addon: sample-based quan- tum diagonalization,”https://github.com/Qiskit/ qiskit-addon-sqd(2024)

    work page 2024

  40. [41]

    J. Yu, J. R. Moreno, J. T. Iosue, L. Bertels, and et al., arXiv (2025), arXiv:2501.09702 [quant-ph]

    work page arXiv 2025

  41. [42]

    Universal Quantum Computer Simulation of 50 Qubits on Europe`s First Exascale Supercomputer Harnessing Its Heterogeneous CPU-GPU Architecture

    H. De Raedt, J. Kraus, A. Herten, V. Mehta, M. Bode, M. Hrywniak, K. Michielsen, and T. Lippert, “Universal quantum simulation of 50 qubits on europe’s first exas- cale supercomputer harnessing its heterogeneous cpu-gpu architecture,” (2025), arXiv:2511.03359 [quant-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  42. [43]

    N. F. Robertson, A. Akhriev, J. Vala, and S. Zhuk, arXiv e-prints (2023), 10.48550/arXiv.2301.08609, arXiv:2301.08609 [quant-ph]

    work page doi:10.48550/arxiv.2301.08609 2023

  43. [44]

    Qelvin – a compact quantum circuit simu- lator,

    D. N¨ asman, “Qelvin – a compact quantum circuit simu- lator,” (2024), accessed 2026-03-09

    work page 2024

  44. [45]

    Statevector simulation — qiskit aer,

    DeepWiki, “Statevector simulation — qiskit aer,” (2025), documentation page, accessed 2026-03-09. 14

    work page 2025

  45. [46]

    Efficient classical simulation of slightly entangled quantum computations

    G. Vidal, Physical Review Letters91, 147902 (2003), arXiv:quant-ph/0301063 [quant-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2003

  46. [47]

    Matrix product state simulation method — qiskit aer,

    Qiskit Contributors, “Matrix product state simulation method — qiskit aer,” (2021), tutorial/documentation page, accessed 2026-03-09

    work page 2021

  47. [48]

    Mps (tebd),

    R. Pfeifer, “Mps (tebd),” (2026), tensors.net tutorial page, accessed 2026-03-09

    work page 2026

  48. [49]

    S.-B. B. Lee, H. R. Choi, D. D. Ohm, and S.- S. B. Lee, Physical Review Research7, 033252 (2025), arXiv:2504.04769 [quant-ph]

    work page arXiv 2025

  49. [50]

    Efficient classical calculation of the Quantum Natural Gradient,

    T. Jones, “Efficient classical calculation of the Quantum Natural Gradient,” (2020), arXiv:2011.02991 [quant-ph]

    work page arXiv 2020

  50. [51]

    Operator-projected variational quantum imag- inary time evolution,

    A. A. Anuar, F. Jamet, F. Gironella, F. S. IV, and R. Rossi, “Operator-projected variational quantum imag- inary time evolution,” (2024), arXiv:2409.12018 [quant- ph]

    work page arXiv 2024

  51. [52]

    Gacon, C

    J. Gacon, C. Zoufal, G. Carleo, and S. Woerner, Quan- tum5, 567 (2021)

    work page 2021

  52. [53]

    Halla, Quantum9, 1798 (2025), arXiv:2502.17231 [quant-ph]

    M. Halla, Quantum9, 1798 (2025), arXiv:2502.17231 [quant-ph]

    work page arXiv 2025

  53. [54]

    Scalability of quantum error mitigation techniques: from utility to advantage,

    S. N. Filippov, S. Maniscalco, and G. Garc´ ıa-P´ erez, “Scalability of quantum error mitigation techniques: from utility to advantage,” (2024), arXiv:2403.13542 [quant-ph]. Appendix A: Alternative Strategies for Estimating the QGT The efficient evaluation of the QGT is one of the main computational bottlenecks in the application of VarQTE. For small prob...

    work page arXiv 2024

  54. [55]

    Among these, stochastic approximation methods are particularly promising because they avoid the quadratic scaling in the number of parameters

    Related W ork: Stochastic Approximation Methods Several strategies have been proposed to reduce the quantum resources required to calculate the QGT. Among these, stochastic approximation methods are particularly promising because they avoid the quadratic scaling in the number of parameters. We focus on three representative approaches. The method developed...

    work page

  55. [56]

    Classical Evaluation of the SPSA Fidelity T erms The SPSA estimator reduces the problem of approximating the QGT to repeated evaluations of the fidelity between a variational state at parametersθ t and nearby states obtained by perturbing those parameters. This observation suggests a possible hybrid strategy: rather than evaluating the fidelities on quant...

    work page

  56. [57]

    To answer this, we analyze the SPSA estimator under assumptions that are deliberately favorable to the method

    Why Stochastic Methods F ail in Practice Given the promising theoretical properties of stochastic methods such as SPSA, a critical question is whether they can achieve the precision required for practical VarQTE within reasonable computational budgets. To answer this, we analyze the SPSA estimator under assumptions that are deliberately favorable to the m...

    work page