Pith. sign in

REVIEW 3 major objections 5 minor 45 references

This paper claims that a fixed short-time evolution under a constant, unit-strength 'solver' Hamiltonian can prepare many-body ground states, with a warm-start-plus-incremental-coupling strategy reaching fidelities ~0.999 for 10-qubit chain

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 ↗

Short-time evolution under an optimized constant two-body solver Hamiltonian can prepare Heisenberg ground states with ~99.9% fidelity, provided warm starts are combined with incremental coupling ramps.

T0 review reviewed 2026-08-04 challenge →

load-bearing objection A promising but overclaimed numerical study of a warm-start combo for analog variational state preparation; the scaling claim exceeds the evidence. the 3 major comments →

arxiv 2511.12923 v2 pith:HYRXHTG3 submitted 2025-11-17 quant-ph

Time-Efficient Quantum Many-Body State Synthesis and its Optimization via Warm Start Strategies

Prashasti Tiwari , Dylan Lewis , Sougato Bose This is my paper
classification quant-ph
keywords ground state preparationquantum many-body systemsvariational quantum eigensolverHamiltonian ansatzwarm startHeisenberg modelanalog quantum simulationstate synthesis
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 asks whether a quantum many-body ground state—normally prepared by slow adiabatic ramping—can instead be generated by a single fast, fixed-time (t=1) evolution under a time-independent two-body 'solver' Hamiltonian whose couplings are of order unity. The solver's parameters are found classically by minimizing the energy of the evolved state with respect to the problem Hamiltonian. The authors demonstrate this for random anisotropic Heisenberg chains up to 10 qubits and complete graphs up to 6 qubits, and identify a scaling strategy: warm-starting from an optimized n−1-qubit solver and then slowly ramping up the couplings of the added qubit. This combination, which they call the combo method, achieves final fidelities of roughly 0.999 (chains) and 0.99 (complete graphs) and avoids the gradient decay that plagues cold starts. If the approach holds at larger sizes, ground states could be prepared much faster than adiabatic methods allow, and the optimized solvers could be catalogued and reused on analog quantum simulators.

Core claim

The central discovery is that a shallow analog variational ansatz—a single time evolution e^{-i H_s t} with t=1 under an anisotropic Heisenberg Hamiltonian H_s with arbitrary local fields and couplings restricted to the same graph as the problem Hamiltonian H_p—can synthesize the ground state of H_p from a product fiducial state with high fidelity. Optimizing H_s by minimizing the energy expectation value ⟨ψ|H_p|ψ⟩ works, and the key practical finding is that among cold starts, warm starts by size expansion, incremental coupling ramps, and their combination, the combined warm-start-plus-ramp strategy is the only one that reliably reaches high fidelity: it yields fidelities ≈0.999 for n=10 ch

What carries the argument

The central object is the solver Hamiltonian H_s, an anisotropic XYZ Heisenberg model on the same interaction graph as the problem Hamiltonian, augmented with site-dependent local fields. Time evolution under H_s for a fixed unit time acts as a single deep entangling gate; because the ansatz is parameterized by a single time-independent Hamiltonian, the recent no-barren-plateau result applies. The essential algorithmic strategy is the 'combo' warm start: first optimize the solver for an n−1-qubit chain, then add a qubit and slowly scale its couplings from small to order-unity values, using the previous parameters as the starting point at each ramp step. A diagnostic metric β = ΔE/||∇C|| trac

Load-bearing premise

The load-bearing premise is that for any random problem Hamiltonian of the considered form, the target ground state lies within the reachable set of states generated by a single unit-time evolution under a solver Hamiltonian with couplings of order unity on the same graph—a reachability property the paper assumes without proof and supports only with numerical evidence on small instances.

What would settle it

Run the same warm-start-combo optimization on an n=12 or n=14 spin chain with random anisotropic couplings and check whether the achieved fidelity stays above, say, 0.99; if fidelity drops sharply with n, the claimed scaling recipe fails. Alternatively, take an optimized H_s from one random H_p and evaluate the energy of the state it produces for a different random H_p; poor transfer would undercut the catalog-and-reuse premise.

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

If this is right

  • If the central claim holds, ground states of small many-body spin systems can be prepared in a fixed time ~1 (in units of inverse coupling), independent of the spectral gap, sidestepping adiabatic slowdown.
  • The optimized solver Hamiltonians are reusable: found once classically, they can be catalogued and programmed into trapped-ion or neutral-atom simulators to rapidly produce the corresponding ground states on demand.
  • The warm-start combo recipe provides a concrete scaling path for variational state preparation that mitigates gradient decay, potentially extending to n≈12–14 qubits with more compute.
  • Because H_s differs from H_p, the method is not a direct Hamiltonian simulation but a fast state-synthesis protocol, useful for initial states in quench dynamics, quantum chemistry, and sensing.
  • The result gives numerical evidence that expressibility of a fixed-time, constant-Hamiltonian evolution is sufficient for random Heisenberg ground states on chain and complete-graph topologies.

Where Pith is reading between the lines

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

  • The paper does not prove reachability; the assumption that e^{-i H_s t} can express the target ground state for arbitrary random H_p remains extrapolative beyond n=10 and n=6. A natural test is to probe fidelity decay with n.
  • The optimized H_s is instance-specific; the authors do not test whether a solver found for one random H_p transfers to another. If transferability holds, the classical amortization argument becomes much stronger.
  • The warm-start combo resembles curriculum learning in variational quantum algorithms; a similar incremental-coupling schedule might accelerate other Hamiltonian variational ansätze beyond the Heisenberg class.
  • The method's dependence on graph matching between H_s and H_p suggests a topology limitation: for highly nonlocal problem Hamiltonians, the same-graph restriction may need to be relaxed, with unknown cost in parameters and optimization difficulty.
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 / 5 minor

Summary. The paper proposes a time-efficient method to prepare ground states of spin Hamiltonians Hp (anisotropic Heisenberg models on a graph) by evolving a simple fiducial state for a fixed short time t=1 under a time-independent 'solver' Hamiltonian Hs of the same graph plus local fields. Hs parameters are optimized classically by minimizing the energy of the evolved state with respect to Hp. Four optimization strategies are compared: cold start, warm start by size expansion, incremental coupling ramp, and a combination ('combo'). Numerical results for 100 random instances of XYZ chains up to n=10 and complete graphs up to n=6 show that the combo method achieves the highest fidelity (~0.999 for chains, ~0.99 for graphs). The authors claim that this combo strategy is the best scaling strategy for preparing n-qubit ground states and speculate that it may extend to n~12–14.

Significance. If the numerical claims hold, this work offers an interesting alternative to digital VQE and adiabatic preparation: a single or double constant-Hamiltonian evolution of unit time, with O(n^2) parameters, can produce high-fidelity ground states, and the warm-start/incremental combination may mitigate barren-plateau issues. The use of exact diagonalization only as a diagnostic, not as part of the cost function, correctly avoids circularity. However, the significance is currently limited: the central scaling claim rests on a small number of sizes, averaged curves without error bars, and no reachability or out-of-sample validation. The paper is better read as a proof-of-concept than as an established scaling law; the evidence is not yet sufficient for the strong claims made in the abstract and conclusion.

major comments (3)
  1. [Section III, Eqs. (2)–(4)] The method's viability depends on the existence of a solver Hamiltonian Hs in the restricted family (same graph as Hp, local fields, Jmax~1, t=1, L≤2) such that e^{-iHs t}|ψ0⟩ approximates |GS(Hp)⟩. The Lie-algebraic motivation in Section I concedes that the coefficients of nested commutators are not independent, yet no proof, bound, or numerical measure of reachability is provided. Please report worst-case and percentile fidelities over the 100 instances, and provide optimized Hs parameters for representative Hp instances so that reachability can be independently checked. Without this, the scaling claim is a conjecture.
  2. [Section IV, Figs. 3–4] The abstract and conclusion claim the combo strategy is 'the best scaling strategy,' and the abstract mentions up to n=14, but the data cover only n=10 chains and n=6 complete graphs. No fidelity-versus-n curves are shown, and the averaged curves in Figs. 3–4 lack error bars or per-instance variance. A single snapshot at one n cannot establish a scaling advantage over other strategies; the authors should present per-n means and variances and compare iteration counts or total computational cost across n to identify a scaling trend. The 'n~12–14' statement in the Conclusion is explicitly speculative and should be clearly labeled as such.
  3. [Section III and IV (reproducibility)] The optimization details are underspecified: L-BFGS convergence criterion, the exact incremental-ramp schedule (number of steps, coupling distribution, Jmax values), initial parameter distributions, and the number of random restarts are not given. No optimized Hs couplings are provided, so the numerical results cannot be reproduced or independently verified. Please supply pseudo-code for the combo method and include at least one representative Hs (or a link to code) in the supplementary material.
minor comments (5)
  1. [Section IV] The sentence 'The metrics for the n=6 complete graph and n=10 chain graph are shown in Fig.3 and Fig.4 respectively' contradicts the figure captions, where Fig.3 is labeled 'N=10 chain' and Fig.4 is 'N=6 complete graph'. Please correct the cross-reference.
  2. [Section V] The Conclusion states the cost function is ⟨GS(Hp)|Hp|ψ(Js)⟩, which would require prior knowledge of the ground state and is inconsistent with Eq. (5) of Section III, where the cost is ⟨ψ(Js)|Hp|ψ(Js)⟩ and the exact ground state is used only as a diagnostic. If this is a typo, please correct it; as written it implies circularity.
  3. [Abstract vs. body] The abstract says 'up to n=14 qubit many-body states', but the body presents n=10 chains and n=6 complete graphs; the conclusion only speculates about n~12–14. Please align the abstract with the actual results.
  4. [Throughout] There are several typographical errors: 'fidelty', 'explot', 'upto', and inconsistent use of n vs. N in Eqs. (3)–(4). Reference [31] contains an unusual DOI ('10.1103/rgyh-8xw8') that appears to be a placeholder.
  5. [Section IV] The text says the simulation 'do[es] not simulate quantum measurements' on the output, but Fig. 2 includes measurement symbols. Clarify that the measurement is only for the intended hardware implementation, not for the classical simulation.

Circularity Check

0 steps flagged

No significant circularity: central result is an empirical optimization comparison with independently computed fidelity.

full rationale

The optimization is self-contained and not circular: the cost function is C(Js) = <psi(Js)|Hp|psi(Js)> (Eq. 5), and the reported fidelity F(Js) (Eq. 8) is measured against |GS(Hp)> obtained by exact numerical diagonalization, an independent reference not used in the optimizer. The comparison of cold start, warm start, incremental ramp, and combination ramp is a direct empirical comparison of convergence behavior; the combo strategy is not defined in terms of the winning metric, so its reported advantage is not true by construction. The only self-citations (Refs. 11, 20, 39) are background/motivation and do not carry the derivation. The paper explicitly labels the solver ansatz as a choice ('taking a bit of academic freedom here in proposing the above Hs as our ansatz') and concedes the reachability limitation ('the coefficients of the terms are not all independent'); these are admitted assumptions rather than hidden circular reductions. Unsupported extrapolation (e.g., 'up to n=14' in the abstract vs. numeric data up to n=10 chains and n=6 complete graphs) is an evidence gap, not a circularity. I therefore find no step in which a claimed prediction reduces by construction to its own inputs.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 0 invented entities

No new physical entities are introduced. The central dependence is on the freely parameterized solver Hamiltonian and on the unproven reachability/smoothness assumptions of the variational landscape. The warm-start strategy is itself a heuristic that the paper's own data shows to be non-trivial: pure warm start fails, combo succeeds, suggesting the landscape structure is more complex than assumed.

free parameters (4)
  • solver Hamiltonian couplings and fields (Hs parameters) = optimized values not reported; only Jmax ~ 1 constraint given
    The entire method depends on classically optimizing all couplings Jx, Jy, Jz and local fields hx, hy, hz of Hs. The paper does not report the found values, so the central 'recipe' cannot be inspected or reused.
  • number of layers L = L=1 for chains, L=2 for complete graphs
    Chosen by hand per topology. The paper does not study whether L=1 is sufficient for graphs or why L=2 is needed.
  • maximum coupling strength Jmax = ~1
    Chosen to make the unit-time evolution 'rapid'. This is a physically motivated but hand-set scale. The paper does not study sensitivity of the results to this value.
  • initial fiducial state choice = |0>^n or |+>^n
    The choice of initial state is a free input. The paper does not report per-instance performance for both, so the dependence is unquantified.
axioms (4)
  • domain assumption The exponential of a sum of non-commuting terms generates a sufficiently large Lie algebra to reach the target ground states.
    Section I argues the ansatz has large reach via nested commutators, but no proof of controllability/reachability is given. This is load-bearing: if the target ground state is not reachable under the restricted Hs, the optimization can never succeed.
  • domain assumption The energy landscape of the analog variational ansatz is smooth enough for L-BFGS to converge to a high-fidelity solution.
    The method relies on gradient-based optimization. Ref [34] is cited for absence of barren plateaus in a single-Hamiltonian setting, but the paper does not connect this theorem to the specific Heisenberg instances or to the warm-start strategies used here.
  • domain assumption A solution optimized for one random Hp instance provides a useful starting point for nearby instances.
    Warm start by size expansion assumes that the optimized parameters for n-1 qubits are a better starting point than random initialization for n qubits. The paper finds this is not true alone, and the combo is needed; this assumption is partially falsified by their own data.
  • standard math Exact classical diagonalization of Hp provides an independent ground truth for fidelity and energy.
    Valid for n<=10, but it means the scaling demonstration is limited to classically simulable sizes and cannot establish behavior at larger n where the method would be needed.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Time-Efficient Quantum Many-Body State Synthesis and its Optimization via Warm Start Strategies." pith.science (2026). https://pith.science/paper/HYRXHTG3

@misc{pith2026251112923,
  author       = {Pith},
  title        = {Pith review of: Time-Efficient Quantum Many-Body State Synthesis and its Optimization via Warm Start Strategies},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HYRXHTG3}},
  note         = {Machine review of arXiv:2511.12923}
}
Share X Bluesky LinkedIn Reddit HN
abstract

Quantum mechanical ground states of many-body systems can be important resources for various investigations: for quantum sensing, for benchmarking quantum hardware with classically solvable states, as the initial states for nonequilibrium quantum dynamics following quenches, the simulation of quantum processes that start by coupling systems in ground states, eg, could be a process in quantum chemistry, while their approximations are required as inputs to quantum phase estimation algorithm. However, preparing ground states can be challenging; for example, it may require adiabatic switching of Hamiltonian terms slower than an inverse gap, which can be time consuming and bring in decoherence. Here we investigate the possibility of preparing a many-body entangled ground state of a certain Hamiltonian, which can be called a quantum ``problem'' Hamiltonian, using the time evolution of an initial fiducial state by another time independent ``solver'' Hamiltonian with couplings up to unit strength for a very short fixed (unit) time: a ``time efficient'' ansatz. The parameters of the solver Hamiltonian are optimised classically minimising energy as the cost function. We present a study of up to $n=14$ qubit many-body states prepared using this methodology. Importantly, we find that a strategy of combining a warm start (an already prepared ground state of a $n-1$ qubit Hamiltonian) and incrementally adding extra couplings of a qubit is the best scaling strategy to prepare the ground state of a $n$-qubit Hamiltonian.

Figures

Figures reproduced from arXiv: 2511.12923 by Dylan Lewis, Prashasti Tiwari, Sougato Bose.

Figure 1
Figure 1. Figure 1: FIG. 1. An illustration of the different proposed methods for [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Analog variational optimisation circuits used in this [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. The data gathered for the for a 6 qubit system, [PITH_FULL_IMAGE:figures/full_fig_p006_4.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

45 extracted references · 2 canonical work pages

  1. [1]

    W arm start by size expansion:Here the parameters optimised for an-qubit system, then system size grows by adding additional qubits, where the previous parameters ofnqubit system form the starting point for then+ 1qubit system

  2. [2]

    and quantum computation/simulation methods [3, 4]. Often the challenge is presented as the requirement to accurately determine the ground state energy; however, for that, one needs to determine the ground state itself accurately enough. One way is to prepare it in a simula- tor/quantum computer and measure its energy [5]. In quantum simulators, it is ofte...

  3. [3]

    Combination Ramp:Here both the techniques 1 and 2 are combined, where the system size is increased, then starting with very low coupling strength with the new qubit, the coupling strength is increased till it reaches a maximumJ max ∼1, where the number of qubits is increased

  4. [4]

    Incremental coupling ramp:Here the optimi- sation begins with weak strength withJ <<1(which is just the shorthand forallJ α ij <<1) and is gradually increased to the full strengthJmax∼1in steps, where the parameters obtained for the previous strength are used for the next round of iterations

  5. [5]

    N. J. Ward, I. Kassal, and A. Aspuru-Guzik, Prepara- tion of many-body states for quantum simulation, The Journal of chemical physics130(2009)

    2009

  6. [6]

    This is kept as a control method to be able to compare to the others

    Cold Start:Here the optimisation starts for the final number of qubitsn, with initialisation|0⟩ ⊗n or |+⟩⊗n and there are no stored parameters. This is kept as a control method to be able to compare to the others. All the Warm start schemes (1, 2 and 3) use a pre- vious solution of a state that is close to the final state, thus explot the smoothness of th...

  7. [7]

    Sachdev,Quantum Phase Transitions, 2nd ed

    S. Sachdev,Quantum Phase Transitions, 2nd ed. (Cam- bridge University Press, 2011)

    2011

  8. [8]

    Orús, A practical introduction to tensor networks: Matrix product states and projected entangled pair states, Annals of Physics349, 117 (2014)

    R. Orús, A practical introduction to tensor networks: Matrix product states and projected entangled pair states, Annals of Physics349, 117 (2014)

    2014

  9. [9]

    R. P. Feynman, Simulating physics with computers, In- ternational Journal of Theoretical Physics21(1982)

    1982

  10. [10]

    Preskill, Quantum computing in the nisq era and be- yond, Quantum2, 79 (2018)

    J. Preskill, Quantum computing in the nisq era and be- yond, Quantum2, 79 (2018)

    2018

  11. [11]

    They may also be useful for studying thermalization/non-thermalization in closed and open systems following quenches with entan- gled ground states as starting states [13]

    or how a system changes due to an external change of environment or internal couplings [12], say, by a sud- den change of pressure on a solid. They may also be useful for studying thermalization/non-thermalization in closed and open systems following quenches with entan- gled ground states as starting states [13]. Those states thatcanbepreparedmay also se...

    Pith/arXiv arXiv 2025

  12. [12]

    Vojta, Quantum phase transitions, Reports on Progress in Physics66, 2069 (2003)

    M. Vojta, Quantum phase transitions, Reports on Progress in Physics66, 2069 (2003)

    2069

  13. [13]

    Greiner, O

    M. Greiner, O. Mandel, T. Esslinger, T. W. Hänsch, and I. Bloch, Quantum phase transition from a superfluid to a mott insulator in a gas of ultracold atoms, Nature415, 39 (2002)

    2002

  14. [14]

    Li, Y.-H

    Y.-C. Li, Y.-H. Zhou, Y. Zhang, and H.-Q. Lin, Low ly- ing excited states quantum entanglement and continuous quantum phase transitions, Scientific Reports15, 6277 (2025), article number: 6277

    2025

  15. [15]

    T. A. Bespalova and O. Kyriienko, Quantum simulation and ground state preparation for the honeycomb kitaev model, arXiv preprint arXiv:2109.13883 (2021)

    Pith/arXiv arXiv 2021

  16. [16]

    R. J. MacDonell, C. E. Dickerson, C. J. Birch, A. Kumar, C. L. Edmunds, M. J. Biercuk, C. Hempel, and I. Kas- sal, Analog quantum simulation of chemical dynamics, Chemical science12, 9794 (2021)

    2021

  17. [17]

    Bayat, S

    A. Bayat, S. Bose, and P. Sodano, Entanglement routers using macroscopic singlets, Physical review letters105, 187204 (2010)

    2010

  18. [18]

    Barmettler, M

    P. Barmettler, M. Punk, V. Gritsev, E. Demler, and E. Altman, Quantum quenches in the anisotropic spin- heisenberg chain: different approaches to many-body dy- namics far from equilibrium, New Journal of Physics12, 055017 (2010)

    2010

  19. [19]

    Sels, Bath-induced delocalization in interacting dis- ordered spin chains, Physical Review B106, L020202 (2022)

    D. Sels, Bath-induced delocalization in interacting dis- ordered spin chains, Physical Review B106, L020202 (2022)

    2022

  20. [20]

    Patkowski, O

    M. Patkowski, O. Ayyildiz, M. Kebrič, K. L. Hunt, and D. Lee, High-fidelity scalable quantum state preparation via the fusion method, arXiv preprint arXiv:2510.19039 (2025)

    Pith/arXiv arXiv 2025

  21. [21]

    Argüello-Luengo, A

    J. Argüello-Luengo, A. González-Tudela, T. Shi, P. Zoller, and J. I. Cirac, Analogue quantum chemistry simulation, Nature574, 215 (2019)

    2019

  22. [22]

    Gibbs, Z

    J. Gibbs, Z. Holmes, and P. Stevenson, Exploiting sym- metries in nuclear hamiltonians for ground state prepa- ration, Quantum Machine Intelligence7, 14 (2025)

    2025

  23. [23]

    M. Umer, E. Mastorakis, S. Evangelou, and D. G. An- gelakis, Probing the limits of variational quantum al- gorithms for nonlinear ground states on real quantum hardware: The effects of noise, Physical Review A111, 012626 (2025)

    2025

  24. [24]

    Jaderberg, G

    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 matrix product states on quantum computers, arXiv preprint arXiv:2503.09683 (2025)

    Pith/arXiv arXiv 2025

  25. [25]

    Lubasch, V

    M. Lubasch, V. Murg, U. Schneider, J. I. Cirac, and M.- C. Banuls, Adiabatic preparation of a heisenberg antifer- romagnet using an optical superlattice, Physical review letters107, 165301 (2011)

    2011

  26. [26]

    Farooq, A

    U. Farooq, A. Bayat, S. Mancini, and S. Bose, Adiabatic many-body state preparation and information transfer in quantum dot arrays, Physical Review B91, 134303 (2015)

    2015

  27. [27]

    Calzetta, Not-quite-free shortcuts to adiabatic- ity, Physical Review A98, 10.1103/physreva.98.032107 (2018)

    E. Calzetta, Not-quite-free shortcuts to adiabatic- ity, Physical Review A98, 10.1103/physreva.98.032107 (2018)

    work page doi:10.1103/physreva.98.032107 2018

  28. [28]

    Bharti, A

    K. Bharti, A. Cervera-Lierta, T. H. Kyaw, T. Haug, S. Alperin-Lea, A. Anand, M. Degroote, H. Heimonen, J. S. Kottmann, T. Menke,et al., Noisy intermediate- scale quantum algorithms, Reviews of Modern Physics 94, 015004 (2022)

    2022

  29. [29]

    Tilly, H

    J. Tilly, H. Chen, S. Cao, D. Picozzi, K. Setia, Y. Li, E. Grant, L. Wossnig, I. Rungger, G. H. Booth,et al., The variational quantum eigensolver: a review of meth- ods and best practices, Physics Reports986, 1 (2022)

    2022

  30. [30]

    Wecker, M

    D. Wecker, M. B. Hastings, and M. Troyer, Progress to- wards practical quantum variational algorithms, Physical Review A92, 042303 (2015)

    2015

  31. [31]

    Hadfield, Z

    S. Hadfield, Z. Wang, B. O’gorman, E. G. Rieffel, D. Ven- turelli, and R. Biswas, From the quantum approximate optimization algorithm to a quantum alternating opera- tor ansatz, Algorithms12, 34 (2019)

    2019

  32. [32]

    Kokail, C

    C. Kokail, C. Maier, R. van Bijnen, T. Brydges, M. K. Joshi, P. Jurcevic, C. A. Muschik, P. Silvi, R. Blatt, C. F. Roos,et al., Self-verifying variational quantum simula- tion of lattice models, Nature569, 355 (2019)

    2019

  33. [33]

    C. Lyu, V. Montenegro, and A. Bayat, Accelerated varia- tionalalgorithmsfordigitalquantumsimulationofmany- body ground states, Quantum4, 324 (2020)

    2020

  34. [34]

    Feulner and M

    V. Feulner and M. J. Hartmann, Variational quantum eigensolver ansatz for the j 1-j 2-model, Physical Review B106, 144426 (2022)

    2022

  35. [35]

    H.R.Grimsley, G.S.Barron, E.Barnes, S.E.Economou, and N. J. Mayhall, Adaptive, problem-tailored varia- tional quantum eigensolver mitigates rough parameter landscapes and barren plateaus, npj Quantum Informa- tion9, 19 (2023)

    2023

  36. [36]

    Larocca, S

    M. Larocca, S. Thanasilp, S. Wang, K. Sharma, J. Bia- monte, P. J. Coles, L. Cincio, J. R. McClean, Z. Holmes, and M. Cerezo, Barren plateaus in variational quantum computing, Nature Reviews Physics , 1 (2025)

    2025

  37. [37]

    Bärligea, B

    A. Bärligea, B. Poggel, and J. M. Lorenz, Scalability challenges in variational quantum optimization under stochastic noise, Physical Review A112, 10.1103/rgyh- 8xw8 (2025)

    work page doi:10.1103/rgyh- 2025

  38. [38]

    Korpas, V

    G. Korpas, V. Kungurtsev, and J. Mareček, Undecidable problemsassociatedwithvariationalquantumalgorithms (2025), arXiv:2503.23723 [quant-ph]

    Pith/arXiv arXiv 2025

  39. [39]

    Y. H. Teoh, M. Drygala, R. G. Melko, and R. Islam, Machine learning design of a trapped-ion quantum spin simulator (2019), arXiv:1910.02496 [quant-ph]

    Pith/arXiv arXiv 2019

  40. [40]

    Park and N

    C.-Y. Park and N. Killoran, Hamiltonian variational ansatzwithoutbarrenplateaus,Quantum8,1239(2024)

    2024

  41. [41]

    M. S. Rudolph, J. Miller, D. Motlagh, J. Chen, A. Acharya, and A. Perdomo-Ortiz, Synergistic pretrain- ing of parametrized quantum circuits via tensor net- works, Nature Communications14, 8367 (2023). 9

    2023

  42. [42]

    Dborin, F

    J. Dborin, F. Barratt, V. Wimalaweera, L. Wright, and A. G. Green, Matrix product state pre-training for quan- tum machine learning, Quantum Science and Technology 7, 035014 (2022)

    2022

  43. [43]

    A. A. Mele, G. B. Mbeng, G. E. Santoro, M. Collura, and P. Torta, Avoiding barren plateaus via transferability of smooth solutions in a hamiltonian variational ansatz, Physical Review A106, L060401 (2022)

    2022

  44. [44]

    R. Puig, M. Drudis, S. Thanasilp, and Z. Holmes, Varia- tional quantum simulation: a case study for understand- ing warm starts, PRX Quantum6, 010317 (2025)

    2025

  45. [45]

    Lewis, R

    D. Lewis, R. Wiersema, J. Carrasquilla, and S. Bose, Geodesic algorithm for unitary gate design with time- independent hamiltonians, Phys. Rev. A111, 052618 (2025)

    2025

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.