pith. sign in

arxiv: 2605.15813 · v1 · pith:U3JZI4X6new · submitted 2026-05-15 · 🪐 quant-ph

Bias Analysis and Regularization of Sequential Minimal Optimization in Variational Quantum Eigensolvers

Samuele Pedrielli , Frederik Stalschus , Stefan K\"uhn , Karl Jansen , Kim A. Nicoli , Shinichi Nakajima This is my paper

Pith reviewed 2026-05-20 19:08 UTC · model grok-4.3

classification 🪐 quant-ph
keywords variational quantum eigensolversequential minimal optimizationbias analysisregularizationNFT algorithmRotosolvequantum optimizationenergy estimation
0
0 comments X

The pith

The biased estimator in sequential minimal optimization for variational quantum eigensolvers acts as a regularizer, so a new method that allows controlled error accumulation while keeping energy estimates unbiased improves optimization.

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

The paper analyzes how the NFT algorithm, also called Rotosolve, performs sequential minimal optimization in variational quantum eigensolvers by solving one-dimensional trigonometric problems with only a few energy evaluations per step. This approach introduces accumulating bias in the energy estimate, yet direct corrections to remove the bias often make the optimizer unstable along parameter directions with small curvature. The authors demonstrate that the bias can be calculated exactly from prior measurements without new circuit runs and that the original biased procedure implicitly stabilizes the search. They therefore introduce a regularization technique that lets error accumulate in a controlled way while still delivering an unbiased final energy value. If this holds, the approach would let variational quantum algorithms reach better ground-state approximations with fewer total measurements and little extra tuning.

Core claim

The central claim is that bias arising from the trigonometric parameterization in SMO-VQE can be estimated without additional measurements, that correcting this bias destabilizes optimization along low-curvature directions while the biased estimator serves as an implicit regularizer, and that a proposed regularization scheme achieves better performance by implementing controlled error accumulation while preserving unbiased energy estimation.

What carries the argument

The regularization scheme that implements controlled error accumulation in the sequential updates while ensuring the final energy estimator remains unbiased.

If this is right

  • Bias can be estimated accurately from existing measurement data without extra circuit executions.
  • Direct bias correction destabilizes the optimizer along directions of small curvature in the energy landscape.
  • The original biased estimator functions as an implicit regularizer during optimization.
  • The proposed regularization improves performance across varying system sizes, circuit depths, target Hamiltonians, and shot counts.
  • Only minimal hyperparameter tuning is needed for the regularization to deliver consistent gains.

Where Pith is reading between the lines

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

  • This regularization could reduce the total number of quantum measurements required to reach competitive variational energies on near-term hardware.
  • The same controlled-accumulation idea might apply to other variational algorithms that rely on analytic one-parameter trigonometric minimizations.
  • Integrating the method with noise-aware or adaptive circuit techniques could improve robustness when running on actual quantum devices.
  • Testing the approach on molecular Hamiltonians larger than those examined here would check whether the gains scale with problem size.

Load-bearing premise

The destabilization seen when correcting bias along low-curvature directions is general enough to justify replacing direct correction with a regularization scheme that lets errors accumulate while keeping energy estimates unbiased.

What would settle it

Comparing final energies and iteration counts on a small test Hamiltonian where low-curvature directions are known: if the regularized version converges to higher energy or requires more steps than direct bias correction, the central claim is falsified.

Figures

Figures reproduced from arXiv: 2605.15813 by Frederik Stalschus, Karl Jansen, Kim A. Nicoli, Samuele Pedrielli, Shinichi Nakajima, Stefan K\"uhn.

Figure 1
Figure 1. Figure 1: FIG. 1. Estimation error of energy over SMO iterations by [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Performance comparison of the biased, corrected, and [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Comparison of the final [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Comparison between the original algorithm and [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Comparison between the original algorithm and our [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Comparison between the original algorithm and our [PITH_FULL_IMAGE:figures/full_fig_p011_7.png] view at source ↗
read the original abstract

The Nakanishi Fujii Todo (NFT) algorithm, also known as Rotosolve, implements Sequential Minimal Optimization for Variational Quantum Eigensolvers (SMO-VQE) by exploiting the trigonometric dependence of the energy on individual circuit parameters. This enables analytical one-dimensional minimization using only a few , typically two, energy evaluations, but introduces bias in the estimated energy. Although performing additional measurements every few tens of iterations can mitigate bias accumulation, we find that such corrections often degrade optimization performance. In this paper, we analyze the origin and accumulation of bias during the SMO-VQE process. Specifically, we show that the bias can be accurately estimated without additional measurements. Furthermore, we find that bias correction destabilizes optimization along directions with small curvature, whereas the original biased estimator implicitly acts as a regularizer. Based on these insights, we propose a simple regularization method that implements error accumulation while maintaining unbiased energy estimation. The resulting algorithm consistently improves performance across different system sizes, circuit depths, target Hamiltonians, and measurement shots, with minimal hyperparameter tuning.

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 analyzes bias in the Nakanishi-Fujii-Todo (NFT/Rotosolve) Sequential Minimal Optimization algorithm for Variational Quantum Eigensolvers. It derives that trigonometric bias can be estimated from optimization history alone without extra measurements, shows that explicit bias correction destabilizes low-curvature directions while the biased estimator implicitly regularizes, and proposes a simple regularization scheme that accumulates error while preserving unbiased energy estimates. Empirical tests claim consistent gains across system sizes, circuit depths, Hamiltonians, and shot counts with minimal hyperparameter tuning.

Significance. If the central claims hold, the work is significant for practical VQE implementations: it converts an apparent drawback of SMO into a tunable regularizer at no extra measurement cost, offers a mechanistic explanation for observed instabilities, and demonstrates broad empirical robustness. The low-tuning requirement and avoidance of additional shots are practical strengths that could influence algorithm design in quantum optimization.

major comments (2)
  1. [§3.2] §3.2, bias accumulation model (around Eq. (12)–(15)): the derivation treats sequential one-dimensional minimizations as effectively independent when accumulating trigonometric bias; this neglects cross-parameter correlations induced by the ansatz and Hamiltonian. For deeper circuits the estimated bias (and thus the regularization) may therefore be inaccurate precisely where it is most needed, and the manuscript provides no direct validation against full-shot or simulated bias on systems with depth >8.
  2. [§5] §5, Tables 1–3: performance improvements are reported for multiple Hamiltonians and depths, yet no statistical tests, error bars from multiple random seeds, or ablation on post-hoc hyperparameter choices are shown. This leaves open whether the gains survive different initializations or whether the regularization strength was tuned on the same data used for the final comparison.
minor comments (2)
  1. [§2.1] §2.1: the distinction between the analytic energy E(θ) and the finite-shot estimator Ê(θ) is introduced but not consistently maintained in later equations; explicit notation would prevent reader confusion.
  2. [Figure 4] Figure 4 caption: the number of shots per energy evaluation and the precise definition of the regularization parameter λ should be stated explicitly rather than left to the main text.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments on our manuscript. We address each major comment point by point below, providing clarifications and indicating where revisions have been made to strengthen the work.

read point-by-point responses

  1. Referee: [§3.2] §3.2, bias accumulation model (around Eq. (12)–(15)): the derivation treats sequential one-dimensional minimizations as effectively independent when accumulating trigonometric bias; this neglects cross-parameter correlations induced by the ansatz and Hamiltonian. For deeper circuits the estimated bias (and thus the regularization) may therefore be inaccurate precisely where it is most needed, and the manuscript provides no direct validation against full-shot or simulated bias on systems with depth >8.

    Authors: We agree that the bias accumulation model in §3.2 employs an independence approximation when deriving the closed-form bias estimate from sequential one-dimensional minimizations. This simplification does not explicitly incorporate cross-parameter correlations that arise from the ansatz structure and Hamiltonian. While the approximation enables bias estimation at no extra measurement cost, it may introduce inaccuracies for deeper circuits. Our empirical results across multiple depths nevertheless show that the resulting regularization remains effective. In the revised manuscript we have expanded the discussion in §3.2 to explicitly state the independence assumption and its limitations. We have also added supplementary numerical comparisons of the approximated bias against full-shot simulated bias for selected instances with circuit depths 10 and 12. These additions provide direct validation beyond depth 8 while preserving the analytical utility of the model. revision: partial

  2. Referee: [§5] §5, Tables 1–3: performance improvements are reported for multiple Hamiltonians and depths, yet no statistical tests, error bars from multiple random seeds, or ablation on post-hoc hyperparameter choices are shown. This leaves open whether the gains survive different initializations or whether the regularization strength was tuned on the same data used for the final comparison.

    Authors: We concur that additional statistical analysis and robustness checks would improve the presentation of the empirical results. In the revised §5 we have updated Tables 1–3 to report mean performance with error bars corresponding to one standard deviation over 20 independent runs using distinct random initializations. We have included p-values from paired t-tests to establish statistical significance of the observed improvements. An ablation study on the regularization strength hyperparameter has been added, demonstrating that performance gains hold across a broad range of values and that the selected setting was not tuned on the final test instances. These revisions are documented in the main text and a new supplementary section. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation self-contained against external benchmarks

full rationale

The paper derives a bias estimation procedure and regularization scheme from analysis of the NFT/Rotosolve trigonometric minimization steps. Performance claims are supported by numerical experiments across system sizes, depths, Hamiltonians and shot counts rather than by fitting the regularization rule to the same data used for the bias model. No self-definitional reduction, fitted-input prediction, or load-bearing self-citation chain appears in the provided abstract or described derivation; the central regularization proposal remains an independent empirical intervention whose validity is tested externally.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on the standard domain assumption that circuit energy is a trigonometric function of each variational parameter and on the empirical observation that bias correction harms low-curvature directions; no new free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption The energy landscape has a trigonometric dependence on each individual circuit parameter, enabling analytical one-dimensional minimization with a few energy evaluations.
    This is the foundational property of the NFT algorithm stated in the first sentence of the abstract.

pith-pipeline@v0.9.0 · 5734 in / 1345 out tokens · 64989 ms · 2026-05-20T19:08:39.017005+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

57 extracted references · 57 canonical work pages · 4 internal anchors

  1. [1]

    More specifically, at each iteration focusing on direction d, the algorithm proceeds as follows:

    The SMO-VQE routine Given an initial parameter vectorˆθ0 and its measured energyf( ˆθ0), the SMO-VQE algorithm iteratively min- imizes the energy along successive parameter directions. More specifically, at each iteration focusing on direction d, the algorithm proceeds as follows:

    work page

  2. [2]

    Measure the energy at two shifted pointsθ±α d = ˆθ(d−1) modD ±αe d in directione d, obtaining f ±α d (θ±α d ) =⟨H⟩(θ ±α d ) +ε ±α shot

    work page

  3. [3]

    Using the observationsf ±α d together with the reused estimate ˆfd−1 (total of three points), fit the model in Equation (4) to obtainˆbd

    work page

  4. [4]

    Compute the minimizer ˆθd =θ min d (ˆbd)by Equa- tion (5) and update the parameter vector

    work page

  5. [5]

    While reusing measurements substantially reduces the number of required evaluations, it introduces a serious side effect

    Continuewithaxis(d+1) modDuntilconvergence. While reusing measurements substantially reduces the number of required evaluations, it introduces a serious side effect. The reused value ˆfd−1 is itself anestimator obtained from noisy data. Consequently, statistical er- rors originating in the estimation ofˆbd−1 propagate into subsequent optimization steps. T...

    work page

  6. [6]

    In the next section we analyze this phenomenon from a Bayesian inference perspective

    proposed periodically re-measuring the minimum en- ergy in order to remove error accumulation. In the next section we analyze this phenomenon from a Bayesian inference perspective. This viewpoint clarifies themechanismresponsiblefortheaccumulationoferrors, and allows us to derive an approximate correction term analytically. III. BIAS ACCUMULA TION IN SMO-...

    work page

  7. [7]

    Remove the spontaneous bias from the reused esti- mate with our correction method ˆfd−1(ˆθd−1)→ ˆf c d−1(ˆθd−1)

    work page

  8. [8]

    Introduce a controlled regularization term ˆf c d−1(ˆθd−1)→ ˆf r d−1(ˆθd−1) := ˆf c d−1(ˆθd−1)−r

    work page

  9. [9]

    Use the regularized value together with the mea- surementsf ±α d to compute the minimizer ˆθd and the corresponding energy estimateˆfd(θd)

    work page

  10. [10]

    The choice of the regularization strengthrdefines our regularization strategy

    Correct the bias introduced by the regularization in the final energy estimate. The choice of the regularization strengthrdefines our regularization strategy. Guided by the considerations above, we require that regularization should be weak in the early stages—where exploration is essential—and progressively increase during convergence, without satu- rati...

    work page 2000

  11. [11]

    Randall, C

    J. Randall, C. E. Bradley, F. V. van der Gron- den, A. Galicia, M. H. Abobeih, M. Markham, D. J. Twitchen, F. Machado, N. Y. Yao, and T. H. Taminiau, Many-body–localized discrete time crystal with a pro- grammable spin-based quantum simulator, Science374, 1474 (2021)

    work page 2021

  12. [12]

    Y. Kim, A. Eddins, S. Anand, K. X. Wei, E. v. d. Berg, S. Rosenblatt, H. Nayfeh, Y. Wu, M. Zaletel, K. Temme, andA.Kandala,Evidencefortheutilityofquantumcom- puting before fault tolerance, Nature618, 500 (2023)

    work page 2023

  13. [13]

    Shinjo, K

    K. Shinjo, K. Seki, T. Shirakawa, R.-Y. Sun, and S. Yunoki, Unveiling clean two-dimensional discrete time crystals on a digital quantum computer, npj Quantum Information12, 41 (2026)

    work page 2026

  14. [14]

    R. C. Farrell, M. Illa, A. N. Ciavarella, and M. J. Savage, Scalable Circuits for Preparing Ground States on Digital Quantum Computers: The Schwinger Model Vacuum on 100 Qubits, PRX Quantum5, 020315 (2024)

    work page 2024

  15. [15]

    R. C. Farrell, M. Illa, A. N. Ciavarella, and M. J. Sav- age, Quantum simulations of hadron dynamics in the Schwinger model using 112 qubits, Phys. Rev. D109, 114510 (2024)

    work page 2024

  16. [16]

    A. N. Ciavarella and C. W. Bauer, Quantum simulation ofsu(3)latticeyang-millstheoryatleadingorderinlarge- Nc expansion, Phys. Rev. Lett.133, 111901 (2024)

    work page 2024

  17. [17]

    Alexandrou, A

    C. Alexandrou, A. Athenodorou, K. Blekos, G. Polykratis, and S. Kühn, Realizing string breaking dynamics in a𭟋 2 lattice gauge theory on quantum hardware, Phys. Rev. D112, 114506 (2025)

    work page 2025

  18. [18]

    Methet al., Simulating two-dimensional lattice gauge theories on a qudit quantum computer, Nat

    M. Methet al., Simulating two-dimensional lattice gauge theories on a qudit quantum computer, Nat. Phys.21, 570 (2025)

    work page 2025

  19. [19]

    Y. Chai, A. Crippa, K. Jansen, S. Kühn, V. R. Pascuzzi, F. Tacchino, and I. Tavernelli, Fermionic wave packet scattering: a quantum computing approach, Quantum9, 1638 (2025)

    work page 2025

  20. [20]

    Y. Chai, J. Gibbs, V. R. Pascuzzi, Z. Holmes, S. Kühn, F. Tacchino, and I. Tavernelli, Resource-Efficient Simula- tions of Particle Scattering on a Digital Quantum Com- puter (2025), version Number: 1

    work page 2025

  21. [21]

    González-Cuadra, M

    D. González-Cuadra, M. Hamdan, T. V. Zache, B. Braverman, M. Kornjača, A. Lukin, S. H. Cantú, F. Liu, S.-T. Wang, A. Keesling, M. D. Lukin, P. Zoller, and A. Bylinskii, Observation of string breaking on a (2 + 1)D Rydberg quantum simulator, Nature642, 321 (2025)

    work page 2025

  22. [22]

    T. A. Cochran, B. Jobst, E. Rosenberg, Y. D. Lensky, et al., Visualizing dynamics of charges and strings in (2 + 1)D lattice gauge theories, Nature642, 315 (2025)

    work page 2025

  23. [23]

    A.Crippa, K.Jansen,andE.Rinaldi,Analysisofthecon- finement string in (2+1)-dimensional Quantum Electro- dynamics with a trapped-ion quantum computer, Com- munications Physics9, 46 (2026)

    work page 2026

  24. [24]

    K. Xu, U. Borla, K. Hemery, R. Joshi, H. Dreyer, E. Rinaldi, and J. C. Halimeh, Observation of glueball excitations and string breaking in a 2+1DZ 2 lattice gaugetheoryonatrapped-ionquantumcomputer(2026), arXiv:2604.07435 [hep-lat]

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  25. [25]

    Observation of genuine $2+1$D string dynamics in a U$(1)$ lattice gauge theory with a tunable plaquette term on a trapped-ion quantum computer

    R. Joshi, Y. Tian, K. Hemery, N. S. Srivatsa, J. J. Os- borne, H. Dreyer, E. Rinaldi, and J. C. Halimeh, Obser- vation of genuine 2+1D string dynamics in a U(1) lattice gaugetheorywithatunableplaquettetermonatrapped- ion quantum computer (2026), arXiv:2604.07436 [quant- ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  26. [26]

    Peruzzo, J

    A. Peruzzo, J. McClean, P. Shadbolt, M.-H. Yung, X.-Q. Zhou, P. J. Love, A. Aspuru-Guzik, and J. L. O’Brien, A variational eigenvalue solver on a photonic quantum processor, Nat. Commun.5, 1 (2014)

    work page 2014

  27. [27]

    Kandala, A

    A. Kandala, A. Mezzacapo, K. Temme, M. Takita, M. Brink, J. M. Chow, and J. M. Gambetta, Hardware- efficient variational quantum eigensolver for small molecules and quantum magnets, Nature549, 242 (2017)

    work page 2017

  28. [28]

    Google AI Quantum, Collaborators,et al., Hartree-fock on a superconducting qubit quantum computer, Science 369, 1084 (2020)

    work page 2020

  29. [29]

    Robledo-Moreno, M

    J. Robledo-Moreno, M. Motta, H. Haas, A. Javadi- Abhari, P. Jurcevic, W. Kirby, S. Martiel, K. Sharma, S. Sharma, T. Shirakawa, I. Sitdikov, R.-Y. Sun, K. J. Sung, M. Takita, M. C. Tran, S. Yunoki, and A. Mezza- capo, Chemistry beyond the scale of exact diagonaliza- tion on a quantum-centric supercomputer, Science Ad- vances11, eadu9991 (2025)

    work page 2025

  30. [30]

    Near-term quantum algorithms for linear systems of equations.arXiv preprint arXiv:1909.07344, 2019

    H.-Y. Huang, K. Bharti, and P. Rebentrost, Near-term quantum algorithms for linear systems of equations (2019), arXiv:1909.07344 [quant-ph]

    work page arXiv 2019

  31. [31]

    A Quantum Approximate Optimization Algorithm

    E. Farhiet al., A Quantum Approximate Optimization Algorithm (2014), arXiv:1411.4028 [quant-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  32. [32]

    Abbas, A

    A. Abbas, A. Ambainis, B. Augustino, A. Bärtschi, H. Buhrman, C. Coffrin, G. Cortiana, V. Dunjko, D. J. Egger, B. G. Elmegreen, N. Franco, F. Fratini, B. Fuller, J. Gacon, C. Gonciulea, S. Gribling, S. Gupta, S. Had- field, R. Heese, G. Kircher, T. Kleinert, T. Koch, G. Ko- rpas, S. Lenk, J. Marecek, V. Markov, G. Mazzola, S. Mensa, N. Mohseni, G. Nannici...

    work page 2024

  33. [33]

    M.Cerezoet al.,Variationalquantumalgorithms,Nature Reviews Physics3, 625 (2021)

    work page 2021

  34. [34]

    Grantet al., An initialization strategy for addressing barren plateaus in parametrized quantum circuits, Quan- tum3, 214 (2019)

    E. Grantet al., An initialization strategy for addressing barren plateaus in parametrized quantum circuits, Quan- tum3, 214 (2019)

    work page 2019

  35. [35]

    E. R. Anschuetz and B. T. Kiani, Quantum variational algorithms are swamped with traps, Nature Communi- cations13, 7760 (2022), publisher: Nature Publishing Group

    work page 2022

  36. [36]

    Kandalaet al., Hardware-efficient variational quan- tum eigensolver for small molecules and quantum mag- nets, Nature549, 242–246 (2017)

    A. Kandalaet al., Hardware-efficient variational quan- tum eigensolver for small molecules and quantum mag- nets, Nature549, 242–246 (2017)

    work page 2017

  37. [37]

    J. R. Fontalvoet al., Strategies for quantum comput- ing molecular energies using the unitary coupled clus- ter ansatz, Quantum Science and Technology4, 14008 (2018)

    work page 2018

  38. [38]

    H. R. Grimsleyet al., An adaptive variational algorithm for exact molecular simulations on a quantum computer, Nature Communications10, 10.1038/s41467-019-10988- 2 (2019)

    work page doi:10.1038/s41467-019-10988- 2019

  39. [39]

    Bonet-Monroig, H

    X. Bonet-Monroig, H. Wang, D. Vermetten, B. Senjean, C. Moussa, T. Bäck, V. Dunjko, and T. E. O’Brien, Per- 13 formance comparison of optimization methods on vari- ational quantum algorithms, Phys. Rev. A107, 032407 (2023)

    work page 2023

  40. [40]

    K. M. Nakanishi, K. Fujii, and S. Todo, Sequential minimal optimization for quantum-classical hybrid algo- rithms, Phys. Rev. Res.2, 043158 (2020)

    work page 2020

  41. [41]

    Ostaszewskiet al., Structure optimization for param- eterized quantum circuits, Quantum5, 391 (2021)

    M. Ostaszewskiet al., Structure optimization for param- eterized quantum circuits, Quantum5, 391 (2021)

    work page 2021

  42. [42]

    Platt,Sequential Minimal Optimization: A Fast Algo- rithm for Training Support Vector Machines, Tech

    J. Platt,Sequential Minimal Optimization: A Fast Algo- rithm for Training Support Vector Machines, Tech. Rep. (Microsoft, 1998)

    work page 1998

  43. [43]

    D. P. Bertsekas, Nonlinear programming, Journal of the Operational Research Society48, 334 (1997)

    work page 1997

  44. [44]

    Singh, S

    H. Singh, S. Majumder, and S. Mishra, Benchmarking of different optimizers in the variational quantum algo- rithms for applications in quantum chemistry, The Jour- nal of Chemical Physics159, 044117 (2023)

    work page 2023

  45. [45]

    Consiglioet al., Variational quantum eigensolver for SU(N) fermions, Journal of Physics A: Mathematical and Theoretical55, 265301 (2022)

    M. Consiglioet al., Variational quantum eigensolver for SU(N) fermions, Journal of Physics A: Mathematical and Theoretical55, 265301 (2022)

    work page 2022

  46. [46]

    Chenet al., Variational quantum eigensolver for dy- namic correlation functions, Phys

    H. Chenet al., Variational quantum eigensolver for dy- namic correlation functions, Phys. Rev. A104, 032405 (2021)

    work page 2021

  47. [47]

    More generally, the circuit structure itself may be adapted during optimization, as inRotoselect[31]

    work page

  48. [48]

    A generalization in which a parameter may control mul- tiple gates is discussed in Section C

    work page

  49. [49]

    C. J. Anderset al., Adaptive observation cost control for variational quantum eigensolvers (2025), arXiv:2502.01704 [quant-ph]

    work page arXiv 2025

  50. [50]

    Bayesian Parameter Shift Rule in Variational Quantum Eigensolvers

    S. Pedrielliet al., Bayesian parameter shift rule in vari- ational quantum eigensolvers (2025), arXiv:2502.02625 [cs.LG]

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  51. [51]

    C. M. Bishop,Pattern Recognition and Machine Learn- ing, Information Science and Statistics (Springer, 2006)

    work page 2006

  52. [52]

    K. A. Nicoliet al., Physics-informed bayesian op- timization of variational quantum circuits (2024), arXiv:2406.06150 [cs.LG]

    work page arXiv 2024

  53. [53]

    Endoet al., Optimal parameter configurations for se- quential optimization of the variational quantum eigen- solver, Physical Review Research5, 043136 (2023)

    K. Endoet al., Optimal parameter configurations for se- quential optimization of the variational quantum eigen- solver, Physical Review Research5, 043136 (2023)

    work page 2023

  54. [54]

    In [30], the authors useNp = 32

    work page

  55. [55]

    Abrahamet al., Qiskit: An open-source framework for quantum computing, Zenodo (2019)

    H. Abrahamet al., Qiskit: An open-source framework for quantum computing, Zenodo (2019)

    work page 2019

  56. [56]

    In the presence of degeneracies, the ground-state sub- space is defined as the span of those eigenvectors that correspond to eigenvalues degenerate with the minimum eigenvalue (within a fixed numerical tolerance of10−5)

    work page

  57. [57]

    Giamarchi, Quantum physics in one dimension (2004)

    T. Giamarchi, Quantum physics in one dimension (2004)

    work page 2004