Pith. sign in

REVIEW 67 references

A Geometric Theory of Fermion-to-Qubit Encodings

T0 review · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read Fermion-to-qubit encodings carry an intrinsic geometry that mirrors many-body physics.

desk verdict Genuine geometric reformulation, but the 'exact spectral partition' is asserted without a definition of tree- vs site-dominated modes, so the central claim is not checkable. read the letter →

arxiv 2607.14883 v2 pith:5T4MGR5A submitted 2026-07-16 quant-ph cond-mat.str-el

Lakshya Nagpal , Nishith Reen , S.R. Hassan This is my paper
classification quant-phcond-mat.str-el
keywords fermion-to-qubitencodingsBravyi–KitaevtransformationhypergraphLaplacianFiedlereigenvaluespectralpartitionoptimaltransportWassersteindistanceHubbardmodel
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

The paper argues that exact fermion-to-qubit encodings are not merely computational tools: each encoded Hamiltonian defines a weighted hypergraph whose connectivity and coupling distribution evolve with the physical parameters of the underlying model. It introduces the ratio of algebraic connectivities of the kinetic and interaction hypergraphs, shows this ratio follows the exact law ρ(U) = C/(αU), and uses it to identify two geometric universality classes corresponding to tapered and untapered Bravyi–Kitaev encodings. The central structural claim is an exact spectral partition: for every BK hypergraph studied, the interaction Laplacian separates into exactly one quarter 'tree-dominated' and three quarters 'site-dominated' modes, independent of lattice size. A complementary Xia–Bian–Kais representation maps the Hamiltonian to a coupling-space probability measure, and optimal-transport (Wasserstein) distances locate the same interaction regime as double-occupancy changes. If correct, these geometric diagnostics extract interaction-driven reorganization directly from the encoded operator, without diagonalizing the many-body Hilbert space.

What carries the argument

The machinery is a Pauli-term-to-hyperedge map: each non-identity Pauli string becomes a hyperedge on the qubits it touches, weighted by |coefficient|, then clique-expanded to a graph Laplacian. The Fiedler eigenvalue (second-smallest Laplacian eigenvalue) measures global connectivity, and the ratio ρ(U) compares kinetic versus interaction sectors. The binary-tree architecture of the BK encoding is the mechanism claimed to fix the 1:3 spectral partition: a fixed fraction of interaction modes inherit 'tree-dominated' character from the update structure, while the remaining 'site-dominated' modes carry the rest. The XBK route uses the exactly equivalent diagonal Ising form to define a coupling

What would settle it

Take the 4×4 BK Hubbard interaction hypergraph with the paper's clique-expansion weights (Z-leg weight 1, X/Y-leg weight 1/2) and its stated Laplacian, then compute the full spectrum. If any reproducible mode-classification rule — threshold on eigenvector participation, sign-change structure, or mode ancestry in the BK tree — yields a tree-dominated count other than n_q/4 for any lattice size or interaction strength, the central spectral partition is false.

Watch

Extended reading notes

Core claim

The paper's core claim is that the Bravyi–Kitaev transformation induces a hypergraph geometry with a rigid internal organization. For the Hubbard, spinless t–V, Anderson impurity, and Kitaev models, the interaction hypergraph's Laplacian spectrum splits into two families whose relative sizes are exactly N_tree = n_q/4 and N_site = 3n_q/4 for every system treated, from 4×4 to 16×16 lattices. The Fiedler eigenvalue always belongs to the tree-dominated quarter, which the paper identifies as the reason the characteristic interaction scale U* = C/α is nearly size-independent for untapered encodings and why tapered encodings form a separate branch. In the XBK picture, the same reorganization appea

Load-bearing premise

The exact spectral partition rests on the unstated rule by which Laplacian eigenmodes are classified as tree- or site-dominated; the paper groups eigenvalues visually (Appendix A), and without a precise classifier the claimed exact 1:3 ratio is not independently checkable.

Editorial extensions

If this is right

  • The characteristic interaction scale U* can be read off from the encoded Hamiltonian alone, without computing ground states.
  • Untapered BK encodings of the 2D Hubbard model converge to a near-size-independent geometric limit because the lowest mode always samples the tree-dominated quarter.
  • Tapered and untapered encodings are geometrically distinct classes; comparisons between simulations using different tapering schemes should account for this.
  • The Wasserstein maximum in coupling space serves as an independent wavefunction-free marker for interaction-driven reorganization, matching double-occupancy behaviour.
  • The same construction applies uniformly to Hubbard, spinless t–V, Anderson impurity, and Kitaev models, suggesting a model-independent geometric language.

Reading between the lines

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

  • If the 1:3 partition is exact rather than a numerical coincidence, it should be provable from the recursive binary-tree structure of BK update circuits; a combinatorial derivation would be a natural next step.
  • The framework invites extension to other fermion-to-qubit encodings (superfast, segment-based, or low-weight variants): each would carry its own characteristic hypergraph spectrum, and the mode ratio may differ from 1:3.
  • The paper's own caution about three-point finite-size extrapolation suggests the quantitative value U_c/t ≈ 8.87 should be treated as a trend indicator; the universality-class separation is the more durable claim.
  • Because the BK and XBK probes agree where double occupancy changes fastest, a combined geometric order parameter — connecting spectral connectivity and coupling transport — could be tested against standard correlation functions on the same lattices.
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.

Circularity Check

1 steps flagged · score 4.0 of 10

The 'exact competition law' is a definitional scaling identity; the spectral partition is under-specified but not a circular reduction.

  1. other [Section III.A, Eq. (6); Appendix A.3]
    "Since multiplying every edge weight of a weighted graph by a constant scales its Laplacian spectrum by the same constant, the interaction connectivity satisfies λ2(Lint)=αU... Substituting these relations into Eq. (5) immediately yields ρ(U)=C/(αU), which constitutes an exact geometric relation."

    The observable ρ is defined as λ2(Lhop)/λ2(Lint) in Eq. (5). The 'exact geometric competition law' follows solely from the homogeneity of Lint under the global scaling Jα∝U; no independent information about the encoded Hamiltonian enters. The characteristic scale U*=C/α is simply the point where this defined ratio equals 1, so it is a rearrangement of the input definitions and the linear-U weighting, not a testable prediction. Appendix A confirms this: 'since λ2(Lint(U))=slope·U by construction, the crossing is U*=λ2(Lhop)/slope.'

full rationale

Most of the paper is a self-contained numerical exploration: the hypergraph constructed from Pauli strings, the clique-expanded Laplacian, the spectral statistics, and the Wasserstein distances are all computed directly from the encoded Hamiltonians, with no circular dependence on the conclusions. The two universality classes and the 1:3 spectral partition are empirical groupings of the computed spectra; they are not derived by assuming the conclusion. However, Eq. (8) is underdetermined because the manuscript never specifies a deterministic rule for classifying an eigenmode as 'tree-dominated' versus 'site-dominated'; this is a correctness/falsifiability gap rather than a circular reduction. The self-citation to ref. [1] (the Chakrabarti-Hassan-Shankar distance) is peripheral and not load-bearing for the central geometric claims. The one genuinely construction-bound result is Eq. (6): it is a scaling identity that follows immediately from the definition of ρ and the linear-U scaling of the interaction hypergraph, so presenting it as an 'exact geometric law' overstates its content. For this reason the score is moderate (4), not higher.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central quantitative claims rely on ad hoc choices (tapered lattice selection, undefined mode classification, arbitrary ground metric) and on unproven domain assumptions about the meaning of the geometry. No new physical entities are introduced, but the 'tree-dominated' mode category is an invented concept that is never precisely defined.

free parameters (3)
  • Tapered lattice family {2×2, 2×3, 2×4} = 3 sizes
    The universality class for tapered BK is inferred only from these three lattices; adding other sizes could change the finite-size scaling and the claimed class separation.
  • Mode classification rule for tree vs site families = not specified
    The exact N_tree = n_q/4 spectral partition depends on an undefined rule for assigning eigenmodes to families; without a concrete rule the fraction is not a well-defined output.
  • Ground-metric length for coupling-space transport = ℓ(α)_ij = |S_α|/|c_α|
    The Wasserstein distance depends on this arbitrary ad hoc choice of edge length; no principled justification is given for why shorter length should correspond to stronger couplings.
assumptions (4)
  • standard math Standard spectral graph theory: the clique expansion of a hypergraph to a weighted graph preserves the Laplacian's role as a connectivity measure.
    The paper uses this to justify the Fiedler eigenvalue as a geometric observable (Section II.A).
  • domain assumption The BK encoding's binary-tree architecture is the origin of the claimed universal spectral partition.
    The exact n_q/4 fraction is attributed to the binary tree (Section III.D), but this is an interpretation rather than a proven theorem.
  • domain assumption The XBK mapping produces a probability measure over effective Ising couplings that is a meaningful representation for transport analysis.
    The paper assumes the coupling distribution and the Wasserstein distance capture physically relevant reorganization (Section IV).
  • ad hoc to paper The self-cited 'Chakrabarti-Hassan-Shankar' distance (ref [1]) is valid and appropriately used as a fidelity ground truth.
    The paper uses this quantity in computing D_BK and D_XBK without defining it or justifying its validity in this context (Section IV, Fig. 6).

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Geometric Theory of Fermion-to-Qubit Encodings." pith.science (2026). https://pith.science/paper/5T4MGR5A

@misc{pith2026260714883,
  author       = {Pith},
  title        = {Pith review of: A Geometric Theory of Fermion-to-Qubit Encodings},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5T4MGR5A}},
  note         = {Machine review of arXiv:2607.14883}
}
read the original abstract

Exact fermion to qubit transformations are conventionally regarded as algorithmic tools that translate many-body Hamiltonians into qubit representations for quantum simulation. Here we show that they also define intrinsic geometric representations whose structure encodes physically meaningful information beyond spectral equivalence. We develop a geometric framework based on weighted hypergraphs and coupling space representations constructed from the Bravyi--Kitaev (BK) and Xia--Bian--Kais (XBK) encodings. Within the BK representation, we introduce a geometric observable that compares the algebraic connectivities of the kinetic and interaction hypergraphs, derive its exact analytical dependence on interaction strength, and uncover two geometric universality classes together with an exact spectral organization originating from the binary tree architecture of the encoding. The complementary XBK representation describes the evolution of encoded Hamiltonians through probability measures in coupling space, where optimal transport quantifies interaction-driven reorganization independently of the spectral analysis. Applications to the Hubbard, spinless tV , single impurity Anderson, and Kitaev models demonstrate that these connectivity and transport based geometric descriptions consistently capture the structural evolution of encoded quantum Hamiltonians across distinct classes of many-body systems. Our results establish hypergraph geometry as a new framework for understanding fermion-to-qubit encodings,revealing that they serve not only as computational mappings but also as geometric representations of quantum many-body Hamiltonians.

Figures

Figures reproduced from arXiv: 2607.14883 by the authors.

Figure 1
Figure 1. (c) demonstrates that these geometric observ￾ables can be computed efficiently using sparse graph op￾erations. The computational cost grows smoothly with system size and remains many orders of magnitude below that required for direct many-body diagonalization, em￾phasizing that the proposed framework probes structural information contained in the encoded Hamiltonian itself. The existence of two universality classes … view at source ↗
Figure 2
Figure 2. The second family possesses substantially larger normalized eigenvalues and therefore contributes to the lower-U ∗ branch. The persistence of this separation across all lattice sizes demonstrates that the bimodal￾ity is a property of the encoded geometry rather than a consequence of finite-size effects or accidental spectral crossings. This observation immediately raises a more fundamen￾tal question. Does the divisi… view at source ↗
Figure 3
Figure 3. FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6 [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

67 extracted references · 3 canonical work pages

  1. [1]

    Nagpal, A

    L. Nagpal, A. Kumar, and S. Hassan, Scientific Reports 16, 16249 (2026)

    2026

  2. [2]

    R. P. Feynman, International Journal of Theoretical Physics21, 467– 488 (1999)

    1999

  3. [3]

    Lloyd, Science273, 1073–1078 (1996)

    S. Lloyd, Science273, 1073–1078 (1996)

    1996

  4. [4]

    Preskill, Quantum2, 79 (2018)

    J. Preskill, Quantum2, 79 (2018)

    2018

  5. [5]

    M. A. Nielsen and I. L. Chuang,Quantum Computation and Quantum Information, 10th ed. (Cambridge Univer- sity Press, 2010)

    2010

  6. [6]

    Jordan and E

    P. Jordan and E. Wigner, Zeitschrift fur Physik47, 631–651 (1928)

    1928

  7. [7]

    Bravyi and A

    S. Bravyi and A. Kitaev, Annals of Physics298, 210–226 (2002)

    2002

  8. [8]

    J. T. Seeley, M. J. Richard, and P. J. Love, The Journal of Chemical Physics137, 224109 (2012)

    2012

Show all 67 references
  1. [9]

    Aspuru-Guzik, A

    A. Aspuru-Guzik, A. D. Dutoi, P. J. Love, and M. Head- Gordon, Science309, 1704–1707 (2005)

    2005

  2. [10]

    Peruzzo, J

    A. Peruzzo, J. R. McClean, P. Shadbolt, M.-H. Yung, X.- Q. Zhou, P. J. Love, A. Aspuru-Guzik, and J. L. OBrien, Nature Communications5, 4213 (2014)

    2014

  3. [11]

    Y. Cao, J. Romero, J. P. Olson, M. Degroote, P. D. John- son, M. Kieferova, I. D. Kivlichan, T. Menke, B. Per- opadre, N. P. D. Sawaya, S. Sim, L. Veis, and A. Aspuru- Guzik, Chemical Reviews 10.1021/acs.chemrev.8b00803 (2019)

    2019 doi

  4. [12]

    McArdle, S

    S. McArdle, S. Endo, A. Aspuru-Guzik, S. C. Benjamin, and X. Yuan, Quantum computational chemistry (2018)

    2018

  5. [13]

    Havlicek, M

    V. Havlicek, M. Troyer, and J. D. Whitfield, Physical Review A95, 032332 (2017)

    2017

  6. [14]

    Steudtner and S

    M. Steudtner and S. Wehner, Physical Review A99, 022308 (2019)

    2019

  7. [15]

    S. R. White, Physical Review Letters69, 2863–2866 (1992)

    1992

  8. [16]

    S. R. White, Physical Review B48, 10345–10356 (1993)

    1993

  9. [17]

    Schollwock, Annals of Physics326, 96–192 (2011)

    U. Schollwock, Annals of Physics326, 96–192 (2011)

    2011

  10. [18]

    Orus, Annals of Physics 10.1016/j.aop.2014.06.013 (2014)

    R. Orus, Annals of Physics 10.1016/j.aop.2014.06.013 (2014)

    2014 doi

  11. [19]

    Troyer and U.-J

    M. Troyer and U.-J. Wiese, Physical Review Letters94, 170201 (2005)

    2005

  12. [20]

    R. Xia, T. Bian, and S. Kais, The Journal of Physical Chemistry B 10.1021/acs.jpcb.7b10371 (2017)

    2017 doi

  13. [21]

    R. Xia, T. Bian, and S. Kais, The Journal of Chemical Physics154, 014106 (2021)

    2021

  14. [22]

    Lucas, Frontiers in Physics2, 5 (2014)

    A. Lucas, Frontiers in Physics2, 5 (2014)

    2014

  15. [23]

    Glover, G

    F. Glover, G. A. Kochenberger, and Y. Du, A tu- torial on formulating and using qubo models (2018), arXiv:1811.11538 [quant-ph]

    2018 arXiv

  16. [24]

    Boros and P

    E. Boros and P. L. Hammer, Discrete Applied Mathe- matics123, 155– 225 (2002)

    2002

  17. [25]

    I. G. Rosenberg, Cahiers du Centre d’Etudes de Recherche Operationnelle17, 71– 74 (1975)

    1975

  18. [26]

    Ishikawa, IEEE Transactions on Pattern Analysis and Machine Intelligence33, 1234–1249 (2011)

    H. Ishikawa, IEEE Transactions on Pattern Analysis and Machine Intelligence33, 1234–1249 (2011). 14

    2011

  19. [27]

    Dinkelbach, Management Science13, 492– 498 (1967)

    W. Dinkelbach, Management Science13, 492– 498 (1967)

    1967

  20. [28]

    Choi, Quantum Information Processing10, 343–353 (2011)

    V. Choi, Quantum Information Processing10, 343–353 (2011)

    2011

  21. [29]

    Choi, Quantum Information Processing10, 193–209 (2008)

    V. Choi, Quantum Information Processing10, 193–209 (2008)

    2008

  22. [30]

    Bravyi, J

    S. Bravyi, J. M. Gambetta, A. Mezzacapo, and K. Temme, Tapering off qubits to simulate fermionic hamiltonians (2017), arXiv:1701.08213 [quant- ph]

    2017 arXiv

  23. [31]

    Gottesman,Stabilizer Codes and Quantum Error Cor- rection, Ph.D

    D. Gottesman,Stabilizer Codes and Quantum Error Cor- rection, Ph.D. thesis, California Institute of Technology (1997)

    1997

  24. [32]

    Aaronson and D

    S. Aaronson and D. Gottesman, Proceedings of the Royal Society A461, 3473–3482 (2005)

    2005

  25. [33]

    Kadowaki and H

    T. Kadowaki and H. Nishimori, Physical Review E58, 5355–5363 (1998)

    1998

  26. [34]

    A. B. Finnila, M. A. Gomez, C. Sebenik, C. Stenson, and J. D. Doll, Chemical Physics Letters219, 343– 348 (1994)

    1994

  27. [35]

    Albash and D

    T. Albash and D. A. Lidar, Reviews of Modern Physics 90, 015002 (2018)

    2018

  28. [36]

    McGeoch and C

    C. McGeoch and C. Wang, inProceedings of the ACM In- ternational Conference on Computing Frontiers(2013)

    2013

  29. [37]

    Venturelli, S

    D. Venturelli, S. Mandra, S. Knysh, B. OGorman, R. Biswas, and V. Smelyanskiy, Physical Review X5, 031040 (2015)

    2015

  30. [38]

    A. D. King, J. Carrasquilla, J. Raymond, I. Ozfidan, E. Andriyash, A. J. Berkley, M. Reis, T. Lanting, R. Har- ris, E. Altman, and M. H. Amin, Nature560, 456–460 (2018)

    2018

  31. [39]

    T. F. Ronnnow, Z. Wang, J. Job, S. Boixo, S. V. Isakov, D. Wecker, K. M. Svore, and M. Troyer, Science345, 420–424 (2014)

    2014

  32. [40]

    Hukushima and K

    K. Hukushima and K. Nemoto, Journal of the Physical Society of Japan65, 1604–1608 (1996)

    1996

  33. [41]

    Farhi, J

    E. Farhi, J. Goldstone, S. Gutmann, and M. Sipser, Quantum computation by adiabatic evolution (2000), arXiv:quant-ph/0001106 [quant-ph]

    2000 arXiv

  34. [42]

    Farhi, J

    E. Farhi, J. Goldstone, S. Gutmann, J. M. La- pan, A. Lundgren, and D. Preda, Science 10.1126/sci- ence.1057726 (2001)

    2001 doi

  35. [43]

    Roland and N

    J. Roland and N. J. Cerf, Physical Review A65, 042308 (2002)

    2002

  36. [44]

    Teplukhin, B

    A. Teplukhin, B. K. Kendrick, and D. Babikov, Journal of Chemical Theory and Computation 10.1021/acs.jctc.9b00402 (2019)

    2019 doi

  37. [45]

    Teplukhin, B

    A. Teplukhin, B. K. Kendrick, S. Tretiak, and P. A. Dub, Scientific Reports 10.1038/s41598-020-77315-4 (2020)

    2020 doi

  38. [46]

    Teplukhin, B

    A. Teplukhin, B. K. Kendrick, S. M. Mniszewski, Y. Zhang, A. Kumar, C. F. A. Negre, P. M. Anisi- mov, S. Tretiak, and P. A. Dub, Scientific Reports 10.1038/s41598-021-98331- y (2021)

    2021 doi

  39. [47]

    Teplukhin, B

    A. Teplukhin, B. K. Kendrick, S. M. Mniszewski, S. Tretiak, and P. A. Dub, PLOS ONE 10.1371/ jour- nal.pone.0263849 (2022)

    2022

  40. [48]

    Ushijima-Mwesigwa, C

    H. Ushijima-Mwesigwa, C. F. A. Negre, and S. M. Mniszewski, inProceedings of the ACM International Conference on Computing Frontiers(2017)

    2017

  41. [49]

    Boothby, P

    K. Boothby, P. I. Bunyk, J. Raymond, and A. Roy, Next-generation topology of d-wave quantum processors (2020), arXiv preprint

    2020

  42. [50]

    Verstraete and J

    F. Verstraete and J. I. Cirac, Journal of Statistical Me- chanics: Theory and Experiment , P09012 (2005)

    2005

  43. [51]

    R. C. Ball, Physical Review Letters95, 176407 (2005)

    2005

  44. [52]

    Setia and J

    K. Setia and J. D. Whitfield, The Journal of Chemical Physics , 164104 (2018)

    2018

  45. [53]

    Setia, S

    K. Setia, S. Bravyi, A. Mezzacapo, and J. D. Whitfield, Physical Review Research1, 033033 (2019)

    2019

  46. [54]

    Hagge and N

    T. Hagge and N. Wiebe, Error mitigation via error de- tection using generalized superfast encodings (2023)

    2023

  47. [55]

    S. M. Mniszewski, P. A. Dub, S. Tretiak, P. M. Anisimov, Y. Zhang, and C. F. A. Negre, Downfolding the molecular hamiltonian matrix using quantum community detection (2020), arXiv preprint

    2020

  48. [56]

    Motta, C

    M. Motta, C. Sun, A. T. K. Tan, M. J. ORourke, E. Ye, A. J. Minnich, F. G. S. Brandao, and G. K.-L. Chan, Nature Physics 10.1038/s41567-019-0704-4 (2020)

    2020 doi

  49. [57]

    J. R. McClean, J. Romero, R. Babbush, and A. Aspuru- Guzik, New Journal of Physics18, 023023 (2016)

    2016

  50. [58]

    J. R. McClean, N. C. Rubin, I. D. Kivlichan, K. J. Sung, D. S. Steiger, X. Bonet-Monroig, Y. Cao, C. Dai, E. S. Fried, C. Gidney, P. Gokhale, T. Haner, T. S. Hardikar, V. Havlicek, O. Higgott, C. Huang, J. Izaac, Z. Jiang, X. Liu, S. McArdle, M. Neeley, T. E. OBrien, B. OGorma...

    2017

  51. [59]

    Farhi, J

    E. Farhi, J. Goldstone, and S. Gutmann, A quan- tum approximate optimization algorithm (2014), arXiv:1411.4028 [quant-ph]

    2014 arXiv

  52. [60]

    Hauke, H

    P. Hauke, H. G. Katzgraber, W. Lechner, H. Nishimori, and W. D. Oliver, Reports on Progress in Physics83, 054401 (2020). 15 Appendix A: Computational Methods All results are obtained by classical simulation. This appendix specifies the numerical methods and system sizes used t...

    2020

  53. [61]

    Model and encoding The 2D Hubbard model on anL x ×L y lattice is H=−t X ⟨i,j⟩,σ (c† iσcjσ + h.c.) +U X i ni↑ni↓ −µ X i,σ niσ, (A1) with half filling (Ne =L xLy,µ= 0). Comparison models (1D Hubbard, spinlesst–Vchain, SIAM, Kitaev chain) use analogous conventions; the Kitaev cha...

  54. [62]

    T apering Z2 qubit tapering [30] is performed by writing each Pauli string as a symplectic row over GF(2), taking the kernel of the resulting parity-check matrix to obtain inde- pendent symmetry generators, and projecting the Hamil- tonian onto the sector (over all±1 generator...

  55. [63]

    The TABLE II: System sizes used in each figure.n q is the BK qubit count (Laplacian dimension), untapered unless noted

    Hypergraph Laplacian and clique expansion Each qubit is a vertex; each non-identity Pauli term of coefficientc α is a hyperedgeS α over its support. The TABLE II: System sizes used in each figure.n q is the BK qubit count (Laplacian dimension), untapered unless noted. Dataset ...

  56. [64]

    Table II lists exactly the sizes used

    Eigenvalue solves and system sizes All eigensolves reported in the paper are dense (eigvalsh), affordable sincen q never exceeds a few hun- dred in any submitted figure. Table II lists exactly the sizes used. TheUgrid is non-uniform, denser near the crossover; U ∗ is obtained ...

  57. [65]

    Probabilities are truncated to supportp(x)>10 −10 (renormalized)

    Exact diagonalization and W asserstein-1 distance Ground states for the Wasserstein and double- occupancy diagnostics come from exact diagonalization 16 of the tapered, penalty-free Hamiltonian (n q ≤14), us- ing the eigenvector of least eigenvalue. Probabilities are truncated...

  58. [66]

    Finite-size scaling The tapered crossings (U ∗/t= 7.26,9.51,9.89 forN= 4,6,8) are fit toU ∗(L) =U c +a/ √Nsites. With only three points, this is a one-degree-of-freedom linear re- gression, and we report the resulting uncertainty hon- estly:U c/t= 8.87±2.18 (1σ), with a wide c...

  59. [67]

    All computation is determin- istic except where randomness enters (e.g., bootstrap re- sampling), for which a fixed seed is used

    Reproducibility Simulations ran in a fixed Python environment (NumPy, SciPy/HiGHS, a graph library for shortest paths, a fermion-to-qubit transformation package) on in- stitutional HPC resources. All computation is determin- istic except where randomness enters (e.g., bootstra...

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.