Pith. sign in

REVIEW 4 major objections 6 minor 76 references

Krylov-Lie Algebras for Variational Quantum Algorithms: Geometric, Depth-Aware Insights into Expressivity and Trainability

T0 review · 4 major / 6 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read Krylov-Lie groups give a depth-aware, measure-faithful surrogate for variational quantum circuits, with factorially decaying approximation error and exact non-Haar variance formulas.

desk verdict The Krylov–Lie framework is a real addition to finite-depth VQA landscape theory, and the weighted variance formula is worth engaging; but the main approximation theorem is explicitly conditional on an unproven full-rank assumption, and the paper's central counterexample to Ragone et al. sits in an appendix you should read before citing. read the letter →

arxiv 2607.02626 v3 pith:W5M22WYA submitted 2026-07-02 quant-ph math-phmath.MP

Anžej Margeta-Cacace This is my paper
classification quant-phmath-phmath.MP MSC 81P6817B0122E70
keywords Krylov-LiealgebravariationalquantumalgorithmsbarrenplateauslosslandscapeHaarmeasuret-designsKawada-Itôtheoremreachablemanifold
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This thesis aims to replace Haar-random and t-design models of variational quantum algorithm (VQA) loss landscapes with a finite-dimensional, seed-dependent Lie-group construction called a Krylov-Lie group. The central claim is that the reachable manifold of a shallow ansatz can be approximated, on compact full-rank regions, by such a group together with a canonical Radon-Nikodym density that records how the circuit's sampling measure deviates from Haar. From this, the author derives weighted expectation and variance formulas that reduce exactly to the familiar Lie-algebraic Haar formulas when the density is constant, and that isolate all non-Haar effects into explicit correction operators. The thesis also claims that increasing circuit depth does not, in general, force convergence to Haar, and identifies concrete obstructions, while recovering Haar convergence under natural ergodic conditions. The author is candid that the main approximation theorem is conditional on a full-rank comparison-map assumption, and that the factorial error bounds require a BCH-matching property that fails generically for standard (non-prefix) Krylov-Lie algebras.

What carries the argument

The central object is the Krylov-Lie algebra l_Ψ^{(k)}(S): the Lie algebra generated by compressed generators P_Ψ H_j P_Ψ, where P_Ψ projects onto the Krylov subspace spanned by nested commutators of the generator set S acting on a block seed Ψ up to depth k. The corresponding compact represented group K_Ψ^{(k)} (the Krylov-Lie group) provides a finite-dimensional Lie-group surrogate for the reachable manifold. The argument is carried by the canonical comparison map κ_Ψ^{(k)} from the manifold to the group, the Radon-Nikodym density ρ_Ψ^{(k)} expressing the pushforward of the parameter measure relative to Haar measure, and the visible first- and second-moment correction operators V^{(1)}_j(f

What would settle it

Compute the Jacobian of the canonical comparison map for a concrete shallow ansatz (e.g., 2-layer QAOA on a cycle graph with a generic seed) and check whether the full-rank locus has full Lebesgue measure in parameter space; if the rank drops on a set of positive measure, the pushforward acquires a singular component and the Radon-Nikodym density does not exist, invalidating the variance formulas. Additionally, one could test the cycle-graph QAOA counterexample directly to see whether a visible character or irreducible block has averaged eigenvalue exactly 1 for a continuous-support sampling m

Watch

Extended reading notes

Core claim

The paper's core discovery is that VQA reachable manifolds admit numerically robust, geometrically faithful approximations by compact Krylov-Lie groups of controlled dimension, provided a full-rank comparison map exists. Concretely, for a fixed commutator depth k and block seed Ψ, the Krylov-Lie algebra l_Ψ^{(k)} is generated by compressing the circuit generators to the Krylov subspace built from nested commutators acting on the seed, and the corresponding compact group K_Ψ^{(k)} carries normalized Haar measure. The canonical comparison map κ_Ψ^{(k)} from the reachable manifold to this group pushes forward the parameter measure to an absolutely continuous measure with Radon-Nikodym density ρ

Load-bearing premise

The entire machinery—the measure-faithful approximation, the vanishing singular component, and the Radon-Nikodym density—is conditional on the existence of a seed Ψ_0 in a nonempty stratum such that the canonical comparison map κ_Ψ^{(k)} passes a full-rank Jacobian test on each chart of a finite atlas; the paper proves generic maximal Krylov dimension but not generic full-rank of the comparison map, which also depends on the circuit parameterization and its Jacobian.

Editorial extensions

If this is right

  • Shallow VQA landscape statistics can be computed as Haar integrals on a finite-dimensional Krylov-Lie group reweighted by an explicit, computable density, with error bounded by a factorially decaying term in commutator depth.
  • The standard Lie-algebraic Haar variance formula is recovered as the f=0 special case; all non-Haar effects are isolated into explicit correction operators that can be evaluated from the sampling density's Fourier coefficients.
  • Non-Haar densities can reweight the visible sectors of the landscape and may increase the variance relative to the Haar reference, suggesting a concrete mechanism for mitigating barren plateaus at shallow depths.
  • The claim that sufficiently deep circuits must converge to Haar is false in general; convergence holds precisely under observable ergodic conditions (support generating the observable group), as formalized in the paper's observable Kawada-Itô theorem.
  • For BCH-matched Krylov-Lie algebras, concentration inequalities and variance bounds transfer from the surrogate group to the original circuit up to a factorially small threshold shift, giving a quantitative bridge between Lie-group geometry and circuit trainability.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the full-rank comparison-map assumption is not proven generic for typical ansätze, a practical test would be to compute the Jacobian rank of κ_Ψ^{(k)} for concrete shallow circuits (e.g., low-depth QAOA with generic seeds) to see whether the rank-drop locus has positive measure; if it does, the Radon-Nikodym density fails to exist and the variance formula cannot be applied directly.
  • The variance gap formula suggests a design principle: choose parameter distributions whose pushforward under the comparison map has large visible second-moment Fourier blocks in the sectors of interest, effectively importance-sampling the landscape; this is numerically testable on small instances.
  • The observable-group quotient implies that only the projective/unitary-channel image of the circuit matters for moment convergence, so global phases and other invisible degrees of freedom are irrelevant for design-depth thresholds; this could simplify convergence analyses for structured ansätze.
  • The BCH-matching defect for standard KLAs at grade ≥3 means that practical implementations should either use prefix KLAs or enforce the absorption condition P^{(k−1)}⊂K^{(k)}; otherwise the factorial error bounds do not hold, and the approximation quality may degrade qualitatively.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. This manuscript, formatted as a Master's thesis, introduces Krylov–Lie algebras (KLAs) and Krylov–Lie groups (KLGs) as finite-depth, seed-dependent surrogates for the reachable manifolds of variational quantum algorithms. The central technical claim is Theorem 4.2.1: under a full-rank comparison-map hypothesis, the reachable manifold is approximated, on compact full-rank subsets, by a compact KLG, with the pushforward of the parameter measure absolutely continuous with respect to Haar measure and admitting a Radon–Nikodym density. The thesis then derives weighted first- and second-moment formulas (Theorems 5.1.1 and 5.1.4) in which the usual Haar/Lie-algebraic formulas of Ragone et al. are recovered by setting the density to unity, and non-Haar effects are isolated in explicit visible correction operators. It also argues that depth alone need not drive convergence to Haar, gives a counterexample based on cycle-graph QAOA deferred to Appendix A, and formulates an observable Kawada–Itô theorem for moment convergence on a canonical observable quotient group.

Significance. If the conditional approximation theorem could be made unconditional for a reasonably broad class of ansätze, this framework would be a substantial advance: it would supply a depth-aware, finite-dimensional Lie-group model of VQA landscapes, with exact non-Haar variance corrections and a clear reduction to the Haar case. The manuscript has real strengths: the algebraic-geometric stratification of seed-dependent dimensions is carefully developed; the weighted moment formulas are explicit and internally consistent given the imported Haar identity; and the discussion of obstructions to Haar convergence identifies a concrete representation-theoretic mechanism, with the cycle-graph QAOA counterexample providing a falsifiable test. The paper is also commendably honest about several limitations, including the difficulty of constructing dimension-matched charts and the need for Jordan-algebraic extensions. However, the central approximation theorem is conditional on a nonconstructive full-rank premise that is not established for typical ansätze, and the factorial error bounds apply only to a restricted BCH-matched variant. These issues do not invalidate the framework, but they currently li

major comments (4)
  1. [§4.2, Theorem 4.2.1; §3.4] This is the load-bearing gap identified in the reader's stress test, and it lands.
  2. [§4.2.1, Definition 4.2.10, Theorem 4.2.15] This gap affects the quantitative bridge from the KLG surrogate back to the original circuit.
  3. [§5.1.2, Theorem 5.1.4] Please either prove the Haar moment identity in the KLA setting or state it as a named external result.
  4. [§5.3, Remark 5.1.7] This is an interpretation issue rather than a mathematical inconsistency.
minor comments (6)
  1. [Table of Contents] The header 'AN OBSER V ABLE KA W ADA–ITÔ THEOREM' contains stray spacing and capitalization errors; it should be 'An Observable Kawada–Itô Theorem'.
  2. [Equation (3.2)] The coarea formula and its notation (metric Jacobian JΦ, Hausdorff measure H^{m−n}) are used before the quantities are defined. Please add a brief explanation or a reference at first use.
  3. [Definition 3.2.3 and Remark 3.2.4] The convention W0 = I is introduced after the definition says 1≤r≤k. Please clarify that W≤k includes the empty word and W0.
  4. [Remark 3.2.2] The projector formula '1− P_{s−1}^{i=1} |ψ_i⟩⟨ψ_i|' is garbled; it should presumably be I − Σ_{i=1}^{s−1} |ψ_i⟩⟨ψ_i|, with a definition of the projection operator.
  5. [Chapter 6] The notation H̄ is used both for the conjugate Hilbert space and for entrywise complex conjugation of operators. Please unify the notation to avoid ambiguity.
  6. [Lemma 4.1.1] The lemma assumes a finite atlas and uses compactness of M, but compactness is stated only inside the proof. Please include it in the hypotheses or state it explicitly before the lemma.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; central results are conditional on explicit hypotheses, and the only self-citation is minor and non-load-bearing.

full rationale

The derivation chain is self-contained once its explicit hypotheses are granted. Theorem 4.2.1 assumes the existence of a seed Ψ0 passing a full-rank Jacobian test on each chart and then derives absolute continuity and a Radon–Nikodym density by standard submersion/coarea arguments; the rank hypothesis is not the conclusion in disguise, and §3.4's genericity results are not used to prove the circuit-dependent Jacobian part (the paper's own Remark 4.2.3 notes the difficulty of constructing dimension-matched charts). The BCH/Krylov–Lie error bound (Theorem 4.2.15) is conditional on Definition 4.2.10; for the prefix KLA this condition follows from the word-preservation identity of Theorem 3.3.6, which is a structural property of the compression, and the factorial tail is a genuine analytic BCH estimate, not a renaming of the desired error. The variance formula (Theorem 5.1.4) uses the Ragone et al. Haar-sector expression as an external benchmark and adds explicit correction operators; the f=0 reduction (Eq. 5.13) is a legitimate special case, not a fitted parameter called a prediction. The only self-citation is the 'cragged terrains' empirical motivation [13] (§1.2.3, §5.1.7); it is motivational and not load-bearing for any theorem. No uniqueness theorem or ansatz is imported from same-author prior work. The main limitations—the unproven full-rank premise and the non-genericity of BCH matching for standard KLAs (Remark 4.2.12)—are correctness risks, not circularity.

Assumptions & free parameters 3 free parameters · 5 assumptions · 3 invented entities

The framework adds one structural object (KLA/KLG) plus operator-valued bookkeeping (V^(1), V^(2)) and a quotient (G_obs); these are mathematical definitions, so the 'graviton problem' penalty is light. The real ledger burden is the imported Ragone et al. Haar second-moment identity — an external theorem whose normalization must match exactly — and the non-constructive full-rank/BCH-matched hypotheses that gate the approximation theorem.

free parameters (3)
  • commutator depth k
    The Krylov grade is chosen by hand; it sets the size of the surrogate and the claimed factorial error B^{k+1}/(k+1)!, but no prescription maps circuit layer count L to k.
  • block seed (Ψ, size s)
    The seed determines the Krylov subspace dimension and approximation quality; Theorem 4.2.1(6) asserts an optimal seed exists on a compact substratum but gives no construction or algorithm.
  • BCH constants K⋆, γ, A_C^(ψ), B_C = unspecified (existential)
    The factorial error bounds hold 'for some' constants that are never bounded; without values the super-exponential claims are non-quantitative — any target error can be absorbed by enlarging B_C.
assumptions (5)
  • domain assumption Ragone et al. Haar second-moment identity: for O or ρ in the represented Lie algebra, E_Haar[ℓ²_fluc] = Σ_j P_j(ρ)P_j(O)/dim(l_j)
    Imported as 'the familiar Ragone expression' in the proof of Theorem 5.1.4 without derivation; the normalization of the purities must match [37] exactly, which could not be verified against the original here.
  • domain assumption The reachable set M is a smooth manifold of dimension m (with periodic parameter structure quotiented out) with parameter law ν_M = q dvol_M, q ∈ C¹
    Used throughout Theorem 3.1.1 and Theorem 4.2.1; singular loci of the exponential-product parameterization are assumed away by the full-rank locus.
  • ad hoc to paper Existence of a full-rank canonical comparison map for some seed on every chart (rank test passed at least once)
    A hypothesis of Theorem 4.2.1, not proven for generic ansätze; the entire Radon-Nikodym density analysis collapses without it.
  • ad hoc to paper BCH-matchedness of the KLA (Definition 4.2.10)
    Needed for Lemma 4.2.11 and the factorial bounds; the paper shows matching holds for prefix KLAs and for anticommuting-involutive generator sets, but notes it is 'non-generic and unstable' for standard KLAs (Lemma 4.2.11 remark).
  • standard math Standard analytic tools: Haar measure uniqueness, Lie's second/third theorems, Peter-Weyl, determinantal stratification facts, coarea formula, log-Sobolev inequality on compact Lie groups, Kawada-Itô/Stromberg theorem
    Invoked throughout chapters 3–6 with standard references cited ([17, 28, 25, 33, 40]).
invented entities (3)
  • Krylov-Lie algebra/group (KLA/KLG)
    purpose: Finite-depth Lie-theoretic surrogate for the VQA reachable manifold; carries Haar measure and a Radon-Nikodym density for moment computations.
    A new formal object with no experimental instantiation in the paper; its only outputs are the moment/variance formulas evaluated for a chosen seed and depth.
  • Visible correction operators V^(1)_j(f), V^(2)_jk(f)
    purpose: Isolate non-Haar contributions to loss expectation and variance in terms of the centered sampling density f = q−1.
    Well-defined operators computable from ν and K in principle, but the paper provides no computation or measurement of them, and the barren-plateau-mitigation prediction depends on them without demonstrating that optimizers induce favorable f.
  • Observable group G_obs
    purpose: Quotient of the dynamical group by global phases on which t-copy moment coefficients are uniformly dense; the setting in which the Kawada-Itô convergence theorem becomes necessary and sufficient.
    A defined quotient construction, not a physical postulate; it implies moment convergence is governed by the support of the pushforward measure on G_obs, a checkable mathematical claim but not experimentally tested here.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Krylov-Lie Algebras for Variational Quantum Algorithms: Geometric, Depth-Aware Insights into Expressivity and Trainability." pith.science (2026). https://pith.science/paper/W5M22WYA

@misc{pith2026260702626,
  author       = {Pith},
  title        = {Pith review of: Krylov-Lie Algebras for Variational Quantum Algorithms: Geometric, Depth-Aware Insights into Expressivity and Trainability},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W5M22WYA}},
  note         = {Machine review of arXiv:2607.02626}
}
read the original abstract

Variational quantum algorithms (VQAs) are a leading approach to near-term quantum computation, but their utility is limited by barren plateaus and other pathologies in their loss landscapes. Existing landscape theories based on dynamical Lie algebras, Jordan-algebraic Wishart systems, approximate t-designs, and Haar-random circuits are foundational, but they often neglect the finite-depth geometry of realistic ans\"atze and are therefore ill-suited to the shallow-depth regime, where VQAs are poor approximators of 2-designs and trainability is most feasible. This work introduces Krylov algebras, algebraic structures induced by the Krylov span of a finite generator set acting on one or more seed vectors, as a framework for VQA landscape theory. We show that VQA reachable manifolds can be approximated in a numerically robust, geometrically faithful fashion by Krylov-Lie algebras and groups, and that these structures induce canonical invariant measures for computing expectation values and variances under general sampling measures. In particular, we derive weighted non-Haar variance formulas that recover the usual Lie-algebraic Haar formulas as a special case while isolating non-Haar effects into explicit correction terms. We also show that the common heuristic that sufficiently deep circuit ensembles must converge to Haar fails in general without additional hypotheses, identify concrete obstructions to naive Haar convergence, and recover convergence under natural necessary and sufficient ergodic conditions. Lastly, our formulas further imply that non-Haar contributions to landscape statistics may mitigate barren plateaus by reweighting the visible sectors of the loss landscape, suggesting that VQAs may be more trainable than recent literature has posited.

Review history (2 revisions) →

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

76 extracted references · 8 linked inside Pith

  1. [1]

    Fundamentals of

    Grad,. Fundamentals of. Journal of Noncommutative Geometry , publisher=. 2024 , month=Feb, pages=. doi:10.4171/jncg/563 , number=

    doi:10.4171/jncg/563 2024

  2. [2]

    , title =

    Granirer, Edmond E. , title =. Pacific Journal of Mathematics , volume =. 1965 , pages =

    1965

  3. [3]

    Argabright, L. N. , title =. Proceedings of the American Mathematical Society , volume =. 1966 , pages =

    1966

  4. [4]

    , journal =

    Lawson, Jimmie D. , journal =. Polar and

  5. [5]

    Dynkin, E. B. , title =. Mathematics Surveys and Monographs , year =

  6. [6]

    , title =

    Magnus, W. , title =. Communications on Pure and Applied Mathematics , volume =. 1954 , doi =

    1954

  7. [7]

    Childs, A. M. and Su, Y. and Tran, M. C. and Wiebe, N. and Zhu, S. , title =. Quantum , volume =. 2020 , doi =

    2020

  8. [8]

    and Kuwahara, T

    Mori, T. and Kuwahara, T. and Saito, K. , title =. Phys. Rev. Lett. , volume =. 2016 , doi =

    2016

Show all 76 references
  1. [9]

    On the probability distribution on a compact group

    Kawada, Yukiyosi and It. On the probability distribution on a compact group. Proceedings of the Physico-Mathematical Society of Japan. 3rd Series , volume =. 1940 , pages =

    1940

  2. [10]

    Transactions of the American Mathematical Society , volume =

    Stromberg, Karl , title =. Transactions of the American Mathematical Society , volume =. 1960 , pages =

    1960

  3. [11]

    Applebaum, David , title =

  4. [12]

    Harrison Copp and Charlton Li and Anžej Margeta-Cacace and Amy Qiao , title =

  5. [13]

    Dummit and Richard M

    David S. Dummit and Richard M. Foote , title =. 2004 , isbn =

    2004

  6. [14]

    On the dynamical

    Jonathan Allcock and Miklos Santha and Pei Yuan and Shengyu Zhang , year=. On the dynamical. doi:https://doi.org/10.22331/q-2026-05-29-2119 , url=. 2407.12587 , archivePrefix=

    2026 arXiv

  7. [15]

    2021 , eprint=

    A Diffusion Theory For Deep Learning Dynamics: Stochastic Gradient Descent Exponentially Favors Flat Minima , author=. 2021 , eprint=

    2021

  8. [16]

    A note on the

    Heybetkulu Mustafayev , keywords =. A note on the. Statistics & Probability Letters , volume =. 2022 , issn =. doi:https://doi.org/10.1016/j.spl.2021.109261 , url =

    2022

  9. [17]

    2008 , url =

    Assaf Naor , title =. 2008 , url =

    2008

  10. [18]

    V. M. Petrogradsky , title =. Formal Power Series and Algebraic Combinatorics , publisher =. 2000 , pages =

    2000

  11. [19]

    1992 , isbn =

    Joe Harris , title =. 1992 , isbn =

    1992

  12. [20]

    Beltrametti and Ettore Carletti and Dionisio Gallarati and Giacomo Monti Bragadin , title =

    Mauro C. Beltrametti and Ettore Carletti and Dionisio Gallarati and Giacomo Monti Bragadin , title =. 2009 , isbn =

    2009

  13. [21]

    2016 , eprint=

    Real Algebraic Geometry for Geometric Constraints , author=. 2016 , eprint=

    2016

  14. [22]

    Real Algebraic Geometry , publisher =

    Jacek Bochnak and Michel Coste and Marie-Fran. Real Algebraic Geometry , publisher =

  15. [23]

    Hall , title =

    Brian C. Hall , title =. 2015 , doi =

    2015

  16. [24]

    Antoulas , title =

    Athanasios C. Antoulas , title =

  17. [25]

    Yousef Saad , title =

  18. [26]

    Gutknecht and Thomas Schmelzer , title =

    Martin H. Gutknecht and Thomas Schmelzer , title =. Linear Algebra and its Applications , volume =

  19. [27]

    2006 , isbn =

    Isaac Chavel , title =. 2006 , isbn =

    2006

  20. [28]

    Evans and Ronald F

    Lawrence C. Evans and Ronald F. Gariepy , title =. 2015 , doi =

    2015

  21. [29]

    , title =

    Lee, John M. , title =. 2013 , isbn =

    2013

  22. [30]

    , title =

    Folland, Gerald B. , title =. 1995 , series =

    1995

  23. [31]

    , title =

    Folland, Gerald B. , title =. 1999 , isbn =

    1999

  24. [32]

    2014 , archivePrefix=

    O'Meara, Corey Patrick , title =. 2014 , archivePrefix=. 2510.04719 , primaryClass =

    2014

  25. [33]

    Quantum , volume =

    Hamiltonian variational ansatz without barren plateaus , author =. Quantum , volume =. 2024 , month = feb, doi =

    2024

  26. [34]

    Absence of Barren Plateaus in Finite Local-Depth Circuits with Long-Range Entanglement , volume=

    Zhang, Hao-Kai and Liu, Shuo and Zhang, Shi-Xin , year=. Absence of Barren Plateaus in Finite Local-Depth Circuits with Long-Range Entanglement , volume=. Physical Review Letters , publisher=. doi:10.1103/physrevlett.132.150603 , number=

    doi

  27. [35]

    and Coles, Patrick J

    Holmes, Zoë and Sharma, Kunal and Cerezo, M. and Coles, Patrick J. , year=. Connecting Ansatz Expressibility to Gradient Magnitudes and Barren Plateaus , volume=. PRX Quantum , publisher=. doi:10.1103/prxquantum.3.010313 , number=

    doi

  28. [36]

    Random quantum circuits are approximate unitary t -designs in depth O\! (nt^

    Haferkamp, Jonas , journal =. Random quantum circuits are approximate unitary t -designs in depth O\! (nt^. 2022 , month =

    2022

  29. [37]

    Kirillov, Jr., Alexander , title =

  30. [38]

    , title =

    Meckes, Elizabeth S. , title =. 2019 , isbn =

    2019

  31. [39]

    and Meckes, Mark W

    Meckes, Elizabeth S. and Meckes, Mark W. , title =. Electronic Communications in Probability , volume =. 2013 , doi =

    2013

  32. [40]

    and Çetiner, Elif and Zorn, Maximilian and Linnhoff-Popien, Claudia , month = dec, year =

    Stein, Jonas and Blenninger, Jonas and Bucher, David and Eder, Peter J. and Çetiner, Elif and Zorn, Maximilian and Linnhoff-Popien, Claudia , month = dec, year =. doi:10.48550/arXiv.2407.13012 , abstract =

    doi arXiv

  33. [41]

    Chaos: An Interdisciplinary Journal of Nonlinear Science , author =

    Generating symmetric graphs , volume =. Chaos: An Interdisciplinary Journal of Nonlinear Science , author =. 2018 , pages =. doi:10.1063/1.5064375 , abstract =

    2018 doi

  34. [42]

    and Arrasmith, Andrew and Babbush, Ryan and Benjamin, Simon C

    Cerezo, M. and Arrasmith, Andrew and Babbush, Ryan and Benjamin, Simon C. and Endo, Suguru and Fujii, Keisuke and McClean, Jarrod R. and Mitarai, Kosuke and Yuan, Xiao and Cincio, Lukasz and Coles, Patrick J. , year=. Variational quantum algorithms , volume=. Nature Reviews Ph...

    doi

  35. [43]

    Eggensperger, Katharina and Lindauer, Marius and Hutter, Frank , month = may, year =. Neural. doi:10.48550/arXiv.1709.07615 , abstract =

    doi arXiv

  36. [44]

    Characterizing barren plateaus in quantum ansätze with the adjoint representation , volume=

    Fontana, Enrico and Herman, Dylan and Chakrabarti, Shouvanik and Kumar, Niraj and Yalovetzky, Romina and Heredge, Jamie and Sureshbabu, Shree Hari and Pistoia, Marco , year=. Characterizing barren plateaus in quantum ansätze with the adjoint representation , volume=. Nature Co...

    doi

  37. [45]

    2023 , eprint=

    Iterative Quantum Algorithms for Maximum Independent Set: A Tale of Low-Depth Quantum Algorithms , author=. 2023 , eprint=

    2023

  38. [46]

    and Sauvage, Fr\'

    Ragone, Michael and Bakalov, Bojko N. and Sauvage, Fr\'. A. Nature Communications , volume =. 2024 , doi =

    2024

  39. [47]

    2025 , eprint=

    A Unified Theory of Quantum Neural Network Loss Landscapes , author=. 2025 , eprint=

    2025

  40. [48]

    and Bakalov, Bojko N

    Wiersema, Roeland and Kökcü, Efekan and Kemper, Alexander F. and Bakalov, Bojko N. , month = sep, year =. Classification of dynamical. doi:10.48550/arXiv.2309.05690 , abstract =

    doi arXiv

  41. [49]

    Farhi, Edward and Goldstone, Jeffrey and Gutmann, Sam , month = nov, year =. A. doi:10.48550/arXiv.1411.4028 , abstract =

    doi arXiv

  42. [50]

    , month = aug, year =

    Eddy, Duncan and Kochenderfer, Mykel J. , month = aug, year =. A. doi:10.48550/arXiv.2008.08446 , abstract =

    2008 doi arXiv

  43. [51]

    1979 , note=

    Eigenvalue Techniques in Design and Graph Theory , author =. 1979 , note=

    1979

  44. [52]

    and Maru

    Pisanski, T. and Maru

  45. [53]

    and Cincio, Lukasz and McClean, Jarrod R

    Larocca, Martín and Thanasilp, Supanut and Wang, Samson and Sharma, Kunal and Biamonte, Jacob and Coles, Patrick J. and Cincio, Lukasz and McClean, Jarrod R. and Holmes, Zoë and Cerezo, M. , year=. Barren plateaus in variational quantum computing , volume=. Nature Reviews Phys...

    doi

  46. [54]

    and Ghosh, S

    Braunstein, S.L. and Ghosh, S. and Severini, S , month =. The Laplacian of a graph as a density matrix: a basic combinatorial approach to separability of mixed states , url =. 2006 , journal =. doi:10.1007/s00026-006-0289-3 , pages =

    2006 doi

  47. [55]

    and Choi, S

    Huang, H. and Choi, S. and McClean, J.R. and Preskill, J. , month =. The vast world of quantum advantage , url =

  48. [56]

    and Ying, Rex and Leskovec, Jure , title =

    Hamilton, William L. and Ying, Rex and Leskovec, Jure , title =. NIPS , year =

  49. [57]

    Akiba, Takuya and Sano, Shotaro and Yanase, Toshihiko and Ohta, Takeru and Koyama, Masanori , month = jul, year =. Optuna:. Proceedings of the 25th. doi:10.1145/3292500.3330701 , language =

  50. [58]

    Communications of the ACM , author =

    Understanding the empirical hardness of. Communications of the ACM , author =. 2014 , pages =. doi:10.1145/2594413.2594424 , abstract =

    2014

  51. [59]

    Mathematics of Control, Signals, and Systems , author =

    Approximation by superpositions of a sigmoidal function , volume =. Mathematics of Control, Signals, and Systems , author =. 1989 , pages =. doi:10.1007/BF02551274 , language =

    1989 doi

  52. [60]

    Nature , author =

    Deep learning , volume =. Nature , author =. 2015 , pages =. doi:10.1038/nature14539 , language =

    2015 doi

  53. [61]

    Product integration, its history and applications , author =

  54. [62]

    Fast Simulation of High-Depth QAOA Circuits , url =

    Lykov, Danylo and Shaydulin, Ruslan and Sun, Yue and Alexeev, Yuri and Pistoia, Marco , year =. Fast Simulation of High-Depth QAOA Circuits , url =. doi:10.1145/3624062.3624216 , booktitle =

  55. [63]

    Proceedings of the 3rd International Conference on Learning Representations,

    Adam: A Method for Stochastic Optimization , author=. Proceedings of the 3rd International Conference on Learning Representations,. 2015 , url=

    2015

  56. [64]

    Diaz, N. L. and García-Martín, Diego and Kazi, Sujay and Larocca, Martin and Cerezo, M. , month = oct, year =. Showcasing a. doi:10.48550/arXiv.2310.11505 , abstract =

    doi arXiv

  57. [65]

    Quantum , author =

    Diagnosing. Quantum , author =. 2022 , pages =. doi:10.22331/q-2022-09-29-824 , abstract =

    2022 doi

  58. [66]

    Advances in Knowledge Discovery and Data Mining , series=

    Density-Based Clustering Based on Hierarchical Density Estimates , author=. Advances in Knowledge Discovery and Data Mining , series=. 2013 , publisher=. doi:10.1007/978-3-642-37456-2_14 , isbn=

    2013 doi

  59. [67]

    and Varoquaux, G

    Pedregosa, F. and Varoquaux, G. and Gramfort, A. and Michel, V. and Thirion, B. and Grisel, O. and Blondel, M. and Prettenhofer, P. and Weiss, R. and Dubourg, V. and Vanderplas, J. and Passos, A. and Cournapeau, D. and Brucher, M. and Perrot, M. and Duchesnay, E. , journal=. S...

  60. [68]

    , month = jun, year =

    Low, Richard A. , month = jun, year =. Pseudo-randomness and. doi:10.48550/arXiv.1006.5227 , abstract =

    doi arXiv

  61. [69]

    Identifying hard native instances for the maximum independent set problem on neutral atoms quantum processors

    Cazals, Pierre and others. Identifying hard native instances for the maximum independent set problem on neutral atoms quantum processors. 2025

    2025

  62. [70]

    and Strogatz, S

    Watts, D. and Strogatz, S. , title =. doi:https://doi.org/10.1038/30918 , journal =

    doi

  63. [71]

    Science , volume =

    Albert-László Barabási and Réka Albert , title =. Science , volume =. 1999 , doi =. https://www.science.org/doi/pdf/10.1126/science.286.5439.509 , abstract =

    1999 doi

  64. [72]

    Hagberg and Daniel A

    Aric A. Hagberg and Daniel A. Schult and Pieter J. Swart , title =. Proceedings of the 7th Python in Science Conference ( SciPy2008) , G \"a el Varoquaux

  65. [73]

    Adiabatic quantum computation , author =. Rev. Mod. Phys. , volume =. 2018 , month =. doi:10.1103/RevModPhys.90.015002 , url =

    2018 doi

  66. [74]

    Haemers , title =

    Willem H. Haemers , title =

  67. [75]

    Nature Communications , author =

    Cost function dependent barren plateaus in shallow parametrized quantum circuits , volume =. Nature Communications , author =. 2021 , keywords =. doi:10.1038/s41467-021-21728-w , abstract =

    2021 doi

  68. [76]

    Journal of Symbolic Computation , author =

    Constructing the vertex-transitive graphs of order 24 , volume =. Journal of Symbolic Computation , author =. 1989 , pages =. doi:10.1016/S0747-7171(89)80033-1 , abstract =

    1989 doi

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.