Pith. sign in

REVIEW 4 major objections 5 minor 83 references

This paper claims that Nishimori error thresholds can be computed by a single closed-form map from the clean critical coupling, reproducing known numerical results within a percentage point and giving first analytical estimates for several

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 ↗

T0 review · deepseek-v4-flash

2026-08-01 15:34 UTC pith:Z3LVMGHH

load-bearing objection A new local projection gives closed-form threshold estimates, but the R=4 channel choice is calibrated to known numerics and the bulk matching is an admitted ansatz—treat Table I as heuristic, not derivation. the 4 major comments →

arxiv 2607.18374 v1 pith:Z3LVMGHH submitted 2026-07-20 quant-ph cond-mat.dis-nncond-mat.stat-mech

Nishimori Threshold Estimation for Bayesian Inference and mathbb{Z}_q Surface Code Decoding

Rohit Mukherjee , Simon Trebst This is my paper
classification quant-ph cond-mat.dis-nncond-mat.stat-mech MSC 82B2082B2781P7060K3568P30 PACS 05.50.+q03.67.Pp75.10.Nr
keywords Nishimori thresholdquantum error correctionreplica theoryFourier-Walsh projectionsurface codesPotts modelZq clock modelBerezinskii-Kosterlitz-Thouless transition
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 paper tries to establish that the error threshold of a wide class of disordered statistical-mechanics models—random-bond/random-plaquette Ising models, Potts models, and Z_q clock models—can be estimated analytically from the critical point of the clean (disorder-free) model alone. The mechanism is a Fourier–Walsh projection of the exact replicated single-bond weight at four replicas onto the pair-overlap channel, equated to the clean critical coupling. If correct, this provides instant closed-form threshold estimates for decoding problems, including two consecutive transitions for Z_q surface codes with q≥5. The paper argues this points to a deeper structural connection between clean and Nishimori-disordered critical points.

Core claim

For Ising variables the threshold condition is (1/8) log cosh(4η_c) = β_clean^c, with η_c = tanh^{-1}γ_c and p_c=(1-γ_c)/2, where β_clean^c is the critical coupling of the underlying clean model. The same projection, applied to the appropriate pair channel (coincidence for Potts, first harmonic for clock variables), yields thresholds for q≤4 Potts models and, for Z_q clock models with q≥5, two Nishimori temperatures bounding an information-critical phase. The resulting thresholds match numerical benchmarks within about a percentage point and asymptotically satisfy the Gilbert–Varshamov self-dual entropy relation.

What carries the argument

The central object is the R=4 replicated single-bond log-weight, w_R(x_1,...,x_R)=ln cosh(η Σ x_a) for Ising variables (and its Potts/clock analogues), decomposed in the Fourier–Walsh basis on the finite group. The retained order-parameter channel is the symmetric pair-overlap coordinate X_2=Σ_{a<b} x_a x_b; the projection coefficient K_MRP=⟨y,X_2⟩/⟨X_2,X_2⟩ is the effective pair coupling, and the ansatz sets K_MRP(η_c)=β_clean^c. At R=4 this yields (1/8)ln cosh(4η_c) exactly, separated into a tree-level two-replica term and a negative quartic dressing; R=4 is the smallest algebra containing two disjoint pair channels.

Load-bearing premise

The load-bearing premise is the 'one-channel critical ansatz': the disordered transition occurs exactly where the Fourier–Walsh projected single-bond pair coupling equals the clean critical coupling, with four replicas as the chosen minimal algebra—an assumption the paper explicitly labels an ansatz, not a theorem.

What would settle it

A high-precision Monte Carlo determination of the 3D random-bond Ising Nishimori threshold that deviates from the projection's 22.626% by more than the stated uncertainty—or a measurement of the honeycomb-lattice threshold that rules out the cell-projection estimate of 6.91%—would falsify the matching condition. More directly, computing the actual first-harmonic weight at a higher replica number R>4 and showing that the R=4 'sweet spot' is not where the threshold crosses the numerics would undercut the minimality prescription.

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

If this is right

  • Immediate closed-form threshold estimates for any stabilizer code whose decoding maps to Ising/Potts/clock-type disorder, in any dimension where the clean critical point is known.
  • For Z_q surface codes with q≥5, the method predicts two distinct Nishimori temperatures, implying an intermediate information-critical phase—testable in decoherence experiments and simulations.
  • The approximate Gilbert–Varshamov relation ln q ≈ H_q(T_1^*) + H_q(T_2^*) emerges without any duality input, suggesting a hidden self-dual structure of the construction.
  • Exactness on the Bethe lattice and in the large-q limit provides anchor points where the estimate is expected to be precise, not just empirical.

Where Pith is reading between the lines

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

  • If the one-channel ansatz holds, the same projection could be applied to other clean critical theories (e.g., XY model in d=3) to estimate thresholds for other quantum codes; the paper's XY examples hint at this.
  • The two-replica exactness on trees suggests a possible rigorous connection to message-passing/decoding thresholds (Kesten–Stigum); the R=4 dressing might be interpretable as a loop-correction term in a systematic expansion, though the paper does not claim that.
  • The observed GV-relation saturation could be used as a diagnostic: models where the projected thresholds violate the relation may indicate a breakdown of the single-channel ansatz, offering a testable criterion beyond the known Potts q>4 failure.

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

4 major / 5 minor

Summary. The paper introduces a 'minimal-replica projection' (MRP) scheme that aims to estimate Nishimori critical thresholds from the clean critical coupling of the corresponding disorder-free model. For Ising variables, the exact four-replica single-bond log-weight is Fourier–Walsh projected onto the pair-overlap channel, giving K_MRP = (1/8) log cosh(4η), and the Nishimori critical point is located by matching this quantity to the clean critical coupling β_clean^c (Eq. 12). The authors show that on a Bethe lattice the two-replica sector is exact and reproduces the Kesten–Stigum condition (Eq. 14). For hypercubic lattices, the scheme is claimed to reproduce known numerical thresholds of random-bond and random-plaquette Ising models in d=2–5 within a percentage point, and it is extended to Potts models with q≤4 and to Z_q clock models, yielding two Nishimori temperatures for q≥5 that bracket an intermediate information-critical phase. The paper also reports that the projected clock thresholds approximately satisfy the Gilbert–Varshamov entropy relation without any duality input.

Significance. If the central one-channel ansatz is valid, the paper would provide a closed-form, broadly applicable estimator for Nishimori thresholds across many statistical-mechanics models and stabilizer codes, which is currently accessible only through dedicated numerics. The exact Bethe-lattice derivation (Eq. 14), the explicit Fourier–Walsh algebra (Eqs. 10–11), and the large-q clock asymptotics (Appendix B) are rigorous and useful contributions in their own right. The paper is also transparent about its main limitation, explicitly labeling the matching condition an ansatz and the R=4 choice 'a minimality prescription, not a convergence theorem.' However, because R=4 is calibrated to known thresholds and because the honeycomb example shows the method requires a model-dependent choice of projection cell, the predictive status of the method is currently not established beyond the cases to which it has been fitted or post-hoc adjusted. The numerical agreements are suggestive but do not, by themselves, overcome the lack of a derivation or a systematic error estimate.

major comments (4)
  1. [Eq. (12), Sec. 'Minimal-replica projection'] The central matching condition K_MRP(η_c) = β_clean^c is an unproven ansatz for loopy lattices. The exact Bethe-lattice result (Eq. 14) is a two-replica tree statement; the four-replica dressing K_4 in Eq. (11) is not shown to control the finite-dimensional transition. The paper explicitly acknowledges this is an ansatz, but the entire predictive claim rests on it. A derivation from an RG or field-theoretic argument, or at least a concrete error bound for the one-channel truncation, is needed to distinguish a controlled approximation from an empirical coincidence.
  2. [Sec. 'Why R=4?' and Fig. 2] The choice R=4 is selected as the 'sweet spot' where the projected threshold matches existing numerical values (Fig. 2). This makes the agreement in Table I partly a calibration result, not an independent test. The minimality argument (R=4 being the smallest replica algebra containing two disjoint pair channels) is reasonable but does not uniquely select R=4; higher R values over-dress the channel by the paper's own observation. The authors should provide an out-of-sample prediction—made before comparison to known thresholds—or an independent first-principles criterion for R=4.
  3. [Appendix C, Honeycomb lattice] The method is not unique: for the honeycomb lattice, the single-bond projection gives p_c ≈ 4.83%, about 30% below the numerical ≈6.75%, while a cell projection gives 6.91%. There is no general rule for when a single-bond projection should be replaced by a cell projection. This hidden flexibility undermines the claim of a universal 'local map' from the clean critical coupling to the Nishimori threshold. The paper should either state a criterion for choosing the projection region or present the honeycomb result as evidence that the method is not fully specified for arbitrary lattices.
  4. [Table I, 2D RBIM row] For the most precise benchmark, the projection gives p_c = 10.8205%, while the numerical value is 10.92212(4)%. The difference of ~0.10 percentage points is about 2500σ relative to the quoted numerical uncertainty. Although the paper's claim of 'within a percentage point' is technically true, it obscures the fact that the deviation is many orders of magnitude larger than the benchmark precision. This illustrates that the method has uncontrolled error; the error bars in Table I, which only propagate clean-model uncertainties, do not represent the accuracy of the ansatz.
minor comments (5)
  1. [Eq. (2)] The expression 'p_c = 1/(1+e^{2β_Nishimori_c(Z_2)}) = 10.82...%' appears to have a formatting issue; the closing brace and percent sign are mangled. Please fix the typography.
  2. [Table I] The row '4D RPGM' lists p_c = 10.82...% and cites Ref. [22], which is the 2D RBIM reference. It may be correct if the 4D random-plaquette gauge model is dual to the 2D RBIM, but this should be stated explicitly in the table or text; otherwise the reader cannot assess the provenance of the numerical benchmark.
  3. [Abstract and Table I] The metric 'within a percentage point' is unusually loose for modern Monte Carlo benchmarks (e.g., 2D RBIM has error 4×10^{-5}). Please report relative deviations or absolute errors in the table so the reader can gauge the accuracy of the method against the precision of the benchmarks.
  4. [Appendix C, main text mention] The statement 'some 30% off' for the honeycomb single-bond projection should be accompanied by the actual numbers (4.83% vs ≈6.75%) in the main text; the reader should not have to go to the appendix to see the magnitude of the failure.
  5. [Sec. 'Z_q clock models' and Table S5] The variance decomposition (94–96% first-harmonic weight) is a useful diagnostic of the local projection, but it does not justify the global matching to β_clean^c. The text should not imply that local dominance of the first harmonic validates the one-channel critical ansatz.

Circularity Check

1 steps flagged

Partial circularity: R=4 is selected as the 'sweet spot' against numerical thresholds, so the Table I Ising agreement is calibration rather than independent prediction; the honeycomb cell-projection choice adds hidden flexibility.

specific steps
  1. fitted input called prediction [Eq. (12), 'Why R=4?' section, Fig. 2, Table I]
    "We calibrate the map's accuracy against known Ising thresholds [see Tab. I] ... For all the models at hand, summarized in Fig. 2, we can see that R=4 is indeed the 'sweet spot' where the MRP projection intersects the 'true' numerical value. Note that our line of argument above is a minimality prescription, not a convergence theorem."

    The central matching condition (12) is not uniquely derived; it depends on the structural choice R=4. The paper fixes R=4 because it is the 'sweet spot' where the projected threshold matches existing numerical thresholds, and explicitly denies that the choice is a convergence theorem. Table I then lists those same benchmarks as reproduced predictions. The agreement is therefore in-sample calibration: the ansatz's key discrete degree of freedom was chosen using the numbers it is then said to predict. This is not a definitional identity (K_MRP is computed exactly from the replicated bond weight and beta_clean is external), but it converts the central numerical evidence from an independent test into a selection criterion.

full rationale

The Fourier-Walsh projection coefficient K_MRP is derived exactly from the replicated single-bond log-weight, and the clean critical coupling is an external input, so there is no step where the target threshold appears in the definition of the projected coupling. No load-bearing self-citation chain was found: refs. [8] and [12] are used as benchmarks, not as inputs, and the Bethe-lattice/Kesten-Stigum anchor is a standard external result. The main circularity concern is the choice R=4: the paper states it is the 'sweet spot' where the projection intersects the known numerical values and calls the justification 'a minimality prescription, not a convergence theorem'. Thus the Ising and other Table I agreements used to locate R are calibrated, not independent confirmations. The honeycomb appendix (Appendix C) strengthens this concern by showing a 30% error for the single-bond projection and requiring a hand-introduced cell projection to restore agreement, with the paper itself noting 'there is no bound on its error'. Nevertheless, the method has out-of-sample content (Z_q clock q=5,7 and XY 4D thresholds are stated as first estimates), so the circularity is partial rather than total. Score 4.

Axiom & Free-Parameter Ledger

1 free parameters · 5 axioms · 0 invented entities

The central calculation is exact at the single-bond level, but the mapping to a critical point rests on an unproven matching ansatz and an empirically selected replica number. No new physical entities are introduced.

free parameters (1)
  • Replica number R = 4
    Chosen as the smallest algebra containing two disjoint pair channels, but explicitly selected because R=4 is the 'sweet spot' where projections intersect known numerical thresholds (Fig. 2). This is a hand-picked hyperparameter, not derived from a convergence theorem.
axioms (5)
  • ad hoc to paper Criticality occurs when the Fourier–Walsh projected pair-channel coupling equals the clean critical coupling (K_MRP = β_clean^c).
    Called 'our projection criterion/ansatz' in the text; no derivation from first principles, exact only on the Bethe lattice.
  • ad hoc to paper R=4 is the minimal replica count that captures the relevant local channel; higher R over-dresses the pair channel.
    Stated as 'a minimality prescription, not a convergence theorem'; empirically R=4 matches numerics (Fig. 2).
  • domain assumption Nishimori line / Bayes-optimal matching maps decoding thresholds to critical points, with replica limit R→1 and gauge identities.
    Standard Nishimori/replica setup invoked in 'Nishimori physics' and the inference setting.
  • standard math Fourier–Walsh orthogonality and finite-group Fourier transforms on replicated bond weights.
    Used throughout for the projection; this part is rigorous.
  • domain assumption Clean BKT transition inputs for clock models from Monte Carlo (β_c1, β_c2) are correct as external inputs.
    Table S1 uses MC values for q≥5; propagated uncertainty is given, but any bias in the clean inputs shifts the projected thresholds.

pith-pipeline@v1.3.0-alltime-deepseek · 29097 in / 11611 out tokens · 96936 ms · 2026-08-01T15:34:51.010964+00:00 · methodology

0 comments
read the original abstract

In quantum error correction, the error threshold provides essential quantitative guidance for the ability to bring about fault-tolerance through decoding the effects of incoherent noise, weak measurement or inference. However, the numerical value of an error threshold is typically only accessible through large-scale numerical simulations of the underlying noise model. Here we introduce an analytical estimate of error thresholds falling into the Nishimori universality class via a Fourier--Walsh projection scheme that maps the critical point of the underlying disorder-free statistical-mechanics model to the Born-disordered Nishimori critical point. Using a minimal replica theory approach, this closed-form estimate is obtained from a projection of the exact replicated single-bond weight which we find to reproduce (within a percentage point) the known numerical thresholds of random-bond and random-plaquette Ising models / $\mathbb Z_2$ stabilizer codes in spatial dimensions $d=2-5$, and extends to Potts variables with $q\le4$. The main application of our projection scheme is to $\mathbb Z_q$ surface codes, whose decoding problem maps to the disordered $q$-state clock model. For $q\ge5$ the clean clock model has \textit{two} Berezinskii--Kosterlitz--Thouless transitions, which the projection maps to two Nishimori temperatures that bound an intermediate information-critical phase. The resulting threshold values not only accurately agree with recent decohered-$\mathbb Z_q$-toric-code numerics, but are found to satisfy the Gilbert--Varshamov self-dual entropy relation $\ln q \simeq H_q(T_1^\ast)+H_q(T_2^\ast),$ although no duality condition is imposed in the construction. Our approach thereby points to a deeper connection between the clean and Born-disordered models, while allowing for instant analytical estimates of error thresholds for a variety of stabilizer codes.

Figures

Figures reproduced from arXiv: 2607.18374 by Rohit Mukherjee, Simon Trebst.

Figure 1
Figure 1. Figure 1: Nishimori thresholds for Zq clock models. Shown are the two transition temperatures T ∗ 1,2 of the two subsequent Nishimori transitions of the Zq clock model as a function of q. While the upper temperature saturates at T ∗ 1 → 0.3832 (dotted), the lower projected Nishimori temperature is found to follow T ∗ 2 ≃ A/q2 (dashed). The inset shows the GV self duality condition plotted as function of q which, ide… view at source ↗
Figure 2
Figure 2. Figure 2: Threshold as a function of replica number R. For the variety of models of Table I, we plot the Nishimori threshold estimate as a function of Replica count that we use for our Fourier– Walsh projection. Consistently, we find that the R = 4 estimate (highlighted by the gray bars) matches the existing numerical values (dashed horizontal lines). with X0 = 1 and X2s = P 1≤a1<···<a2s≤R xa1 · · · xa2s . At R = 2 … view at source ↗
Figure 3
Figure 3. Figure 3: Nishimori thresholds for q-state Potts models. Compar￾ison of error correction thresholds obtained by the Hashing bound, our MRP estimations and numerical results [27, 64]. The shaded re￾gion q ≥ 5 indicates the regime where the clean Potts transition is first-order, outside the validity of the projection construction. where the paramagnetic fixed point becomes locally unsta￾ble. For a continuous transitio… view at source ↗
Figure 4
Figure 4. Figure 4: Threshold comparison for the 2D surface code. Shown are the estimates obtained from our projection method, large-scale Monte Carlo simulations [22], coinciding CSS-hashing [33] and Nishimori–Nemoto [65] values. The latter two fall onto the same point because both conditions reduce to H2(p) = 1 2 at q = 2. The error bar of the numerical estimate is smaller than the line width. Appendix B: Zq clock models Fo… view at source ↗
Figure 5
Figure 5. Figure 5: Threshold for hexagonal lattice geometry. We plot the error correction threshold across for honeycomb lattice as a function of replica count R that we use for our Fourier–Walsh projection. For the honeycomb lattice one sublattice is therefore traced out before projecting. In the clean input model this is the well known star-triangle transformation [3], which maps the hon￾eycomb critical point β clean c = 1… 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

83 extracted references · 1 canonical work pages

  1. [1]

    H. A. Kramers and G. H. Wannier, Statistics of the Two- Dimensional Ferromagnet, Phys. Rev.60, 252 (1941), parts I and II; see also Phys. Rev.60, 263 (1941)

    1941

  2. [2]

    Elitzur, Impossibility of spontaneously breaking local sym- metries, Phys

    S. Elitzur, Impossibility of spontaneously breaking local sym- metries, Phys. Rev. D12, 3978 (1975)

    1975

  3. [3]

    R. J. Baxter, Exactly solved models in statistical mechanics, in Integrable systems in statistical mechanics (World Scientific,

  4. [4]

    Nishimori, Exact results and critical properties of the Ising model with competing interactions, Journal of Physics C: Solid State Physics13, 4071 (1980)

    H. Nishimori, Exact results and critical properties of the Ising model with competing interactions, Journal of Physics C: Solid State Physics13, 4071 (1980)

    1980

  5. [5]

    Nishimori, Internal Energy, Specific Heat and Correlation Function of the Bond-Random Ising Model, Progress of Theo- retical Physics66, 1169 (1981)

    H. Nishimori, Internal Energy, Specific Heat and Correlation Function of the Bond-Random Ising Model, Progress of Theo- retical Physics66, 1169 (1981)

    1981

  6. [6]

    Nishimori, Statistical Physics of Spin Glasses and Information Processing (Oxford University Press, Oxford, 2001)

    H. Nishimori, Statistical Physics of Spin Glasses and Information Processing (Oxford University Press, Oxford, 2001)

    2001

  7. [7]

    Dennis, A

    E. Dennis, A. Kitaev, A. Landahl, and J. Preskill, Topological quantum memory, Journal of Mathematical Physics43, 4452 (2002)

    2002

  8. [8]

    G.-Y . Zhu, N. Tantivasadakarn, A. Vishwanath, S. Trebst, and R. Verresen, Nishimori’s Cat: Stable Long-Range Entangle- ment from Finite-Depth Unitaries and Weak Measurements, Phys. Rev. Lett.131, 200201 (2023)

    2023

  9. [9]

    J. Y . Lee, W. Ji, Z. Bi, and M. P. A. Fisher, Decod- ing Measurement-Prepared Quantum Phases and Transitions: from Ising model to gauge theory, and beyond (2022), arXiv:2208.11699 [cond-mat.str-el]

    Pith/arXiv arXiv 2022

  10. [10]

    E. H. Chen, G.-Y . Zhu, R. Verresen, A. Seif, E. B¨aumer, D. Lay- den, N. Tantivasadakarn, G. Zhu, S. Sheldon, A. Vishwanath, S. Trebst, and A. Kandala, Nishimori transition across the error threshold for constant-depth quantum circuits, Nature Physics 21, 161 (2025)

    2025

  11. [11]

    Iba, The Nishimori line and Bayesian statistics, J

    Y . Iba, The Nishimori line and Bayesian statistics, J. Phys. A: Math. Gen.32, 3875 (1999)

    1999

  12. [12]

    P ¨utz, S

    M. P ¨utz, S. J. Garratt, H. Nishimori, S. Trebst, and G.-Y . Zhu, Learning Transitions in Classical Ising Models and Deformed Toric Codes, Phys. Rev. Lett.136, 190402 (2026)

    2026

  13. [13]

    Nahum and J

    A. Nahum and J. L. Jacobsen, Bayesian critical points in clas- sical lattice models, Phys. Rev. B112, 235113 (2025)

    2025

  14. [14]

    R. A. Patil, M. P¨utz, S. Trebst, G.-Y . Zhu, and A. W. W. Ludwig, Higher Nishimori Criticality and Exact Results at the Learning Transition of Deformed Toric Codes (2026), arXiv:2604.06324 [cond-mat.stat-mech]

    Pith/arXiv arXiv 2026

  15. [15]

    K. J. Wiese, A. Das, and A. Nahum, Bayesian phase transition for the critical Ising model: Enlarged replica symmetry in the epsilon expansion and in 2D (2026), arXiv:2604.23346 [cond- mat.stat-mech]

    Pith/arXiv arXiv 2026

  16. [16]

    D. J. Thouless, Spin-Glass on a Bethe Lattice, Phys. Rev. Lett. 56, 1082 (1986)

    1986

  17. [17]

    Carlson, J

    J. Carlson, J. Chayes, L. Chayes, J. Sethna, and D. Thouless, Bethe lattice spin glass: the effects of a ferromagnetic bias and external fields. I. Bifurcation analysis, Journal of Statisti- cal Physics61, 987 (1990)

    1990

  18. [18]

    Carlson, J

    J. Carlson, J. Chayes, J. P. Sethna, and D. Thouless, Bethe lat- tice spin glass: the effects of a ferromagnetic bias and exter- nal fields. II. Magnetized spin-glass phase and the de Almeida- Thouless line, Journal of Statistical Physics61, 1069 (1990)

    1990

  19. [19]

    Takeda, T

    K. Takeda, T. Sasamoto, and H. Nishimori, Exact location of the multicritical point for finite-dimensional spin glasses: a conjec- ture, J. Phys. A: Math. Gen.38, 3751 (2005)

    2005

  20. [20]

    Maillard, K

    J.-M. Maillard, K. Nemoto, and H. Nishimori, Symmetry, com- plexity and multicritical point of the two-dimensional spin glass, J. Phys. A: Math. Gen.36, 9799 (2003)

    2003

  21. [21]

    Ohzeki and J

    M. Ohzeki and J. L. Jacobsen, High-precision phase diagram of spin glasses from duality analysis with real-space renormaliza- tion and graph polynomials, J. Phys. A: Math. Gen.48, 095001 (2015)

    2015

  22. [22]

    Wan, X.-D

    Z.-Q. Wan, X.-D. Dai, and G.-Y . Zhu, Revisiting Nishimori multicriticality through the lens of information measures, Phys. Rev. Res.8, 023059 (2026)

    2026

  23. [23]

    Hasenbusch, F

    M. Hasenbusch, F. P. Toldin, A. Pelissetto, and E. Vicari, Magnetic-glassy multicritical behavior of the three-dimensional ±JIsing model, Phys. Rev. B76, 184202 (2007)

    2007

  24. [24]

    T. Ohno, G. Arakawa, I. Ichinose, and T. Matsui, Phase struc- ture of the random-plaquette Z2 gauge model: accuracy thresh- old for a toric quantum memory, Nuclear Physics B697, 462 (2004)

    2004

  25. [25]

    A. K. Hartmann, Ferromagnetic–spin-glass transition in a four-dimensional random-bond Ising model, Phys. Rev. B64, 224430 (2001)

    2001

  26. [26]

    Fisch, High-temperature series for the ±J random-bond Ising model, Phys

    R. Fisch, High-temperature series for the ±J random-bond Ising model, Phys. Rev. B44, 652 (1991)

    1991

  27. [27]

    R. S. Andrist, J. R. Wootton, and H. G. Katzgraber, Error thresholds for Abelian quantum double models: Increasing the bit-flip stability of topological quantum memory, Phys. Rev. A 91, 042331 (2015)

    2015

  28. [28]

    Vijay and J

    A. Vijay and J. Y . Lee, Information Critical Phases under De- 11 coherence (2025), arXiv:2512.22121 [quant-ph]

    arXiv 2025

  29. [29]

    Alba and E

    V . Alba and E. Vicari, Temperature-disorder phase diagram of a three-dimensional gauge-glass model, Phys. Rev. B83, 094203 (2011)

    2011

  30. [30]

    O’Donnell, Analysis of Boolean Functions (Cambridge Uni- versity Press, 2014)

    R. O’Donnell, Analysis of Boolean Functions (Cambridge Uni- versity Press, 2014)

    2014

  31. [31]

    D. S. Stoffer, Walsh-Fourier analysis and its statistical applica- tions, Journal of the American Statistical Association86, 461 (1991)

    1991

  32. [32]

    Mossel and Y

    E. Mossel and Y . Peres, Information flow on trees (2001), arXiv:math/0107033 [math.PR]

    Pith/arXiv arXiv 2001

  33. [33]

    A. R. Calderbank and P. W. Shor, Good quantum error- correcting codes exist, Phys. Rev. A54, 1098 (1996)

    1996

  34. [34]

    Eckstein, B

    F. Eckstein, B. Han, S. Trebst, and G.-Y . Zhu, Robust Telepor- tation of a Surface Code and Cascade of Topological Quantum Phase Transitions, PRX Quantum5, 040313 (2024)

    2024

  35. [35]

    Le Doussal and A

    P. Le Doussal and A. B. Harris, Location of the Ising Spin- Glass Multicritical Point on Nishimori’s Line, Phys. Rev. Lett. 61, 625 (1988)

    1988

  36. [36]

    Le Doussal and A

    P. Le Doussal and A. B. Harris,ϵexpansion for the Nishimori multicritical point of spin glasses, Phys. Rev. B40, 9249 (1989)

    1989

  37. [37]

    Georges, A., Hansel, D., Le Doussal, P., and Maillard, J.M., The replica momenta of a spin-glass and the phase diagram of n-colour Ashkin-Teller models, J. Phys. France48, 1 (1987)

    1987

  38. [38]

    I. A. Gruzberg, N. Read, and A. W. W. Ludwig, Random-bond Ising model in two dimensions: The Nishimori line and super- symmetry, Phys. Rev. B63, 104422 (2001)

    2001

  39. [39]

    Parisi, The overlap in glassy systems (2013), arXiv:1310.5354 [cond-mat.dis-nn]

    G. Parisi, The overlap in glassy systems (2013), arXiv:1310.5354 [cond-mat.dis-nn]

    Pith/arXiv arXiv 2013

  40. [40]

    Franz and G

    S. Franz and G. Parisi, Phase Diagram of Coupled Glassy Sys- tems: A Mean-Field Study, Phys. Rev. Lett.79, 2486 (1997)

    1997

  41. [41]

    I. R. Pimentel and C. D. Dominicis, Spin glass field theory with replica Fourier transforms, J. Phys. A: Math. Theor.47, 455001 (2014)

    2014

  42. [42]

    T. Aspelmeier, An exact relation between free energy fluctu- ations and bond chaos in the Sherrington–Kirkpatrick model, Journal of Statistical Mechanics: Theory and Experiment2008, P04018 (2008)

    2008

  43. [43]

    See Supplemental Material for definitions of the Fourier–Walsh expansion and weighted least-squares projection, full deriva- tions, sector tables, residual analysis, and detailed comparison with decohered-Zq-toric-code numerics

  44. [44]

    sweet spot

    givesp c ≃0.156atq= 3. Sim- ilarly, forq= 4one can calculate the respective local repli- cated bond weight (detailed in Appendix A and the SM [43]) to obtainp c ≃0.193. These values are in good agreement (i.e. within error bars) with existing numerical values [27], pc = 15.6±0.9 %andp c = 18.3±1.3 %, forq= 3,4, re- spectively. Forq >4the clean Potts trans...

    2004

  45. [45]

    F. Y . Wu, The Potts model, Rev. Mod. Phys.54, 235 (1982)

    1982

  46. [46]

    Duminil-Copin, V

    H. Duminil-Copin, V . Sidoravicius, and V . Tassion, Continuity of the phase transition for planar random-cluster and Potts mod- els with1≤q≤4, Communications in Mathematical Physics 349, 47 (2017)

    2017

  47. [47]

    Kitaev, Fault-tolerant quantum computation by anyons, An- nals of Physics303, 2 (2003)

    A. Kitaev, Fault-tolerant quantum computation by anyons, An- nals of Physics303, 2 (2003)

    2003

  48. [48]

    J. V . Jos´e, L. P. Kadanoff, S. Kirkpatrick, and D. R. Nelson, Renormalization, vortices, and symmetry-breaking perturba- tions in the two-dimensional planar model, Phys. Rev. B16, 1217 (1977)

    1977

  49. [49]

    C. M. Lapilli, P. Pfeifer, and C. Wexler, Universality Away from Critical Points in Two-Dimensional Phase Transitions, Phys. Rev. Lett.96, 140603 (2006)

    2006

  50. [50]

    G. Li, K. H. Pai, and Z.-C. Gu, Tensor-network renormalization approach to theq-state clock model, Phys. Rev. Res.4, 023159 (2022)

    2022

  51. [51]

    H. Chen, P. Hou, S. Fang, and Y . Deng, Monte Carlo study of duality and the Berezinskii-Kosterlitz-Thouless phase tran- sitions of the two-dimensionalq-state clock model in flow rep- resentations, Phys. Rev. E106, 024106 (2022)

    2022

  52. [52]

    Steane, Multiple-particle interference and quantum error correction, Proceedings of the Royal Society of London

    A. Steane, Multiple-particle interference and quantum error correction, Proceedings of the Royal Society of London. Se- ries A: Mathematical, Physical and Engineering Sciences452, 2551 (1996)

    1996

  53. [53]

    Sasamoto and H

    T. Sasamoto and H. Nishimori, Phase structure of the random Zq models in 2D, Progress of Theoretical Physics Supplement 157, 82 (2005)

    2005

  54. [54]

    Baity-Jesi, R

    M. Baity-Jesi, R. A. Ba ˜nos, A. Cruz, L. A. Fernandez, J. M. Gil-Narvion, A. Gordillo-Guerrero, D. I ˜niguez, A. Maiorano, F. Mantovani, E. Marinari, V . Martin-Mayor, J. Monforte- Garcia, A. M. n. Sudupe, D. Navarro, G. Parisi, S. Perez- Gaviro, M. Pivanti, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, B. Seoane, A. Tarancon, R. Tripiccione, a...

    2013

  55. [55]

    Aguilar-Janita, V

    M. Aguilar-Janita, V . Martin-Mayor, J. Moreno-Gordo, and J. J. Ruiz-Lorenzo, Evidence of a second-order phase transition in the six-dimensional Ising spin glass in a field, Phys. Rev. E109, 055302 (2024)

    2024

  56. [56]

    [53, 54], as they count the degree of the integer-moment observable, not the replica number of an aver- aged free energy

    We want to emphasize that these are real replicas in the esti- mator sense of Refs. [53, 54], as they count the degree of the integer-moment observable, not the replica number of an aver- aged free energy. On the Nishimori line the record average of any fixed product of thermal averages is performed exactly at integer replica number, as in Eq. (22), and n...

  57. [57]

    J. R. L. de Almeida and D. J. Thouless, Stability of the Sherrington-Kirkpatrick solution of a spin glass model, Journal of Physics A: Mathematical and General11, 983 (1978)

    1978

  58. [58]

    A. J. Bray and M. A. Moore, Replica symmetry and massless modes in the Ising spin glass, Journal of Physics C: Solid State Physics12, 79 (1979)

    1979

  59. [59]

    M ´ezard, G

    M. M ´ezard, G. Parisi, M. A. Virasoro, and D. J. Thouless, Spin glass theory and beyond (1988)

    1988

  60. [60]

    Sherrington and S

    D. Sherrington and S. Kirkpatrick, Solvable model of a spin- glass, Physical review letters35, 1792 (1975)

    1975

  61. [61]

    At the replica-symmetric Sherrington-Kirkpatrick saddle [59], the overlap Hessian is specified by three matrix elements,P,Q, andR, according to whether two replica pairs share two, one, or no indices [56]. The replicon eigenvalue isλrep =P−2Q+R, with degeneracyn(n−3)/2(here,Ris the matrix element andnis the replica count) and hence two independent modes a...

  62. [62]

    Nishimori Threshold Es- timation for Bayesian Inference andZ q Surface Code Decod- ing

    R. Mukherjee and S. Trebst, Data for “Nishimori Threshold Es- timation for Bayesian Inference andZ q Surface Code Decod- ing”, Zenodo 10.5281/zenodo.21416177 (2026)

    work page doi:10.5281/zenodo.21416177 2026

  63. [63]

    Aizenman and J

    M. Aizenman and J. Wehr, Rounding effects of quenched ran- domness on first-order phase transitions, Communications in mathematical physics130, 489 (1990)

    1990

  64. [64]

    Cardy and J

    J. Cardy and J. L. Jacobsen, Critical Behavior of Random-Bond Potts Models, Phys. Rev. Lett.79, 4063 (1997)

    1997

  65. [65]

    Duclos-Cianci and D

    G. Duclos-Cianci and D. Poulin, Kitaev’sZ d-code threshold estimates, Phys. Rev. A87, 062338 (2013)

    2013

  66. [66]

    Nishimori and K

    H. Nishimori and K. Nemoto, Duality and Multicritical Point of Two-Dimensional Spin Glasses, J. Phys. Soc. Jpn.71, 1198 (2002)

    2002

  67. [67]

    Mezard and A

    M. Mezard and A. Montanari, Information, physics, and computation (Oxford University Press, 2009)

    2009

  68. [68]

    C. H. Bennett, D. P. DiVincenzo, J. A. Smolin, and W. K. Woot- 12 ters, Mixed-state entanglement and quantum error correction, Phys. Rev. A54, 3824 (1996)

    1996

  69. [69]

    K. V . Mardia and P. E. Jupp, Directional statistics (John Wiley & Sons, 2009)

    2009

  70. [70]

    Onsager, Crystal Statistics

    L. Onsager, Crystal Statistics. I. A Two-Dimensional Model with an Order-Disorder Transition, Phys. Rev.65, 117 (1944)

    1944

  71. [71]

    A. M. Ferrenberg, J. Xu, and D. P. Landau, Pushing the limits of Monte Carlo simulations for the three-dimensional Ising model, Phys. Rev. E97, 043301 (2018)

    2018

  72. [72]

    F. J. Wegner, Duality in Generalized Ising Models and Phase Transitions without Local Order Parameters, J. Math. Phys.12, 2259 (1971)

    1971

  73. [73]

    Lundow and K

    P. Lundow and K. Markstr ¨om, Revising the universality class of the four-dimensional Ising model, Nuclear Physics B993, 116256 (2023)

    2023

  74. [74]

    Lundow and K

    P. Lundow and K. Markstr ¨om, The discontinuity of the specific heat for the 5D Ising model, Nuclear Physics B895, 305 (2015)

    2015

  75. [75]

    Li, L.-P

    Z.-Q. Li, L.-P. Yang, Z. Y . Xie, H.-H. Tu, H.-J. Liao, and T. Xi- ang, Critical properties of the two-dimensionalq-state clock model, Phys. Rev. E101, 060105(R) (2020)

    2020

  76. [76]

    Hasenbusch, Monte Carlo study of an improved clock model in three dimensions, Phys

    M. Hasenbusch, Monte Carlo study of an improved clock model in three dimensions, Phys. Rev. B100, 224517 (2019)

    2019

  77. [77]

    J.-P. Lv, W. Xu, Y . Sun, K. Chen, and Y . Deng, Finite-size scal- ing of O (n) systems at the upper critical dimensionality, Na- tional Science Review8, nwaa212 (2021). 13 Supplemental Material: Nishimori Threshold Estimation for Bayesian Inference andZ q Surface Code Decoding Rohit Mukherjee and Simon Trebst Institute for Theoretical Physics, University...

    2021

  78. [78]

    The Ising Hamiltonian has the formH=− P ij σiσj

    Ising bond energy measurement/inference problem We now derive the replicated bond Hamiltonian for the Ising measurement problem atβ= 0before applying the Fourier–Walsh projection. The Ising Hamiltonian has the formH=− P ij σiσj. Let the measured bond variable be a noisy binary recordm ij =±1of the true Ising bond variable τij =σ iσj. The measurement model...

    1930

  79. [79]

    The two clock angles are0andπ, so the cosine becomes a sign

    Clockq= 2 Theq= 2clock model is exactly Ising in disguise. The two clock angles are0andπ, so the cosine becomes a sign. This is a useful check that the generalized clock formula reduces to the ordinary Walsh result. 17 Forq= 2,cos[π(m−∆ a)] = (−1)m(−1)∆a. Withx a = (−1) ∆a, y(∆) = ln 1X m=0 exp " J(−1) m X a xa # = ln[2 cosh(J S)], S= X a xa.(S5) The addi...

  80. [80]

    Grouping them by identical logarithms leaves only three groups with nonzero totalX 1 coefficient

    Clockq= 3 Forq= 3, the translation-reduced sum has3 3 = 27config- urations. Grouping them by identical logarithms leaves only three groups with nonzero totalX 1 coefficient. The table be- low is therefore the entire finite Fourier transform, just written compactly. Forq= 3, Eq. (S4) has denominator 3q3 = 81. Only three grouped logarithms have nonzero tota...

Showing first 80 references.