{"id":"9f5db18c-36c9-4e2b-8fd5-c4603d560d13","arxiv_id":"2501.13611","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Monte Carlo simulations of mapped 3D Z(2) x Z(2) gauge theories yield preliminary toric/surface code thresholds of about 0.68%, 6%, and 1.44% for three circuit-level noise models.","lead":"This proceedings paper maps quantum error correction decoding for the toric/surface code onto three-dimensional Z(2) lattice gauge theories and runs Monte Carlo simulations to estimate threshold error rates under circuit-level noise. The preliminary thresholds it reports sit above those of standard decoders, which suggests better decoding algorithms could push practical fault tolerance further.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quoted threshold values are not yet supported: finite-size extrapolation is absent and Sections 2–3 disagree about which noise model Figs. 2–4 represent.","rationale":"The reader's conditional verdict focuses on two fragilities: the error model derivation, deferred to [11], and the lack of infinite-volume extrapolation. I agree partially. The deferred mapping is a normal proceedings practice and can be checked against the companion paper; it is not an internal inconsistency. The more immediately checkable problem is that the numerical thresholds are not anchored to the displayed data. Section 2 discusses p values (0.00852, 0.00682) that do not appear in the figure captions, while Section 3 and the captions assign the figures to a different noise model and a different threshold (0.0144). This is an internal, text-level inconsistency, not merely a disagreement with current consensus. The qualitative statement that statistical-mechanics thresholds can exceed practical decoders such as MWPM may survive, but the specific numbers should not be used until the finite-size and labeling issues are resolved. Since the reader's verdict is already CONDITIONAL, no change to the verdict is needed; the concern strengthens the conditions rather than rejecting the approach.","tokens_in":7893,"tokens_out":13829,"duration_ms":127832,"concrete_test":"Re-run the Monte Carlo analysis for the three noise models on L = 8, 12, 16, 20, 24 (and at least L = 32 for the two candidate thresholds), using Binder cumulant crossings and susceptibility peaks to extract T_c(L), and extrapolate T_c(L) to infinite volume before determining p_c. Separately, regenerate Fig. 2 for the bit-flip model at p = 0.00852 and p = 0.00682 with the Eq. 1 parameters; if the transition survives at p = 0.00852 or vanishes at p = 0.00682 on the largest lattices, the quoted p_c values are contradicted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central numerical output is p_c ≈ 0.00682, 0.06, and 0.0144, but the evidence connecting those numbers to the plotted data is not secure. Section 2, around Figs. 2–4, describes a well-above-threshold example at p = 0.00852 and a near-threshold example at p = 0.00682, and the conclusion quotes p_c ≈ 0.00682 for bit-flip plus syndrome noise. Section 3 and the figure captions, however, state that Figs. 2–4 are for circuit-level noise (Eq. 3), with displayed probabilities p = 2.88 × 10^-5 and p = 2.31 × 10^-2 and threshold p_c ≈ 0.0144. No plot at p = 0.00852 or p = 0.00682 is actually shown, so the threshold 0.00682 is not tied to displayed data. In addition, the paper admits that at p = 0.020 the transition temperature keeps decreasing with volume and the infinite-volume limit may not exist, yet all p_c estimates are read off finite lattices (L ≤ 24) without a finite-size scaling or extrapolation. If the raw data are as ambiguous as the text suggests, the quoted thresholds could shift substantially; if they survive an infinite-volume extrapolation, the claim would be much better supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This proceedings paper maps the problem of decoding Kitaev's toric/surface code under Pauli errors plus syndrome measurement noise onto three-dimensional Z(2) lattice gauge theories, and reports preliminary Monte Carlo estimates of the threshold probabilities: p_c ≈ 0.00682 for bit/phase-flip plus syndrome noise, p_c ≈ 0.06 for uniform depolarizing plus syndrome noise, and p_c ≈ 0.0144 for anisotropic depolarizing circuit-level noise plus syndrome noise. The thermal transition is studied with the Polyakov line, its susceptibility, and the third-order cumulant on lattices up to 24^3 using Metropolis updates with parallel tempering. The manuscript explicitly labels the results as preliminary.","tokens_in":8139,"tokens_out":3179,"duration_ms":29477,"significance":"If the quoted thresholds are correct, they constitute statistical-mechanics estimates for these noise models that exceed thresholds obtained with minimum-weight perfect matching, and they would suggest room for improved decoders for the surface code. The paper has several strengths: it uses a gauge-invariant order parameter, applies parallel tempering to a frustrated disordered system, and honestly flags the preliminary nature of the results. However, the central numerical claims are not yet supported by the displayed data: no statistical uncertainties are reported, no infinite-volume extrapolation is performed, and the text and figures disagree about which noise model is being shown. The significance of the work therefore depends on follow-up evidence that is not contained in this manuscript.","major_comments":[{"comment":"The manuscript text and the figure captions disagree about what is plotted. Section 2 discusses a well-above-threshold example at p = 0.00852 and a near-threshold example at p = 0.00682 for bit-flip plus syndrome noise, but Figs. 2-4 display only p = 2.88e-5 and p = 2.31e-2 and are described in Section 3 and the captions as circuit-level noise with threshold p_c ≈ 0.0144. No figure shows data at p = 0.00682 or p = 0.00852, so the quoted threshold p_c ≈ 0.00682 is not tied to any displayed result. This inconsistency must be resolved, either by adding the missing panels or by removing the unsupported threshold.","section":"Section 2 (around Figs. 2-4) and Section 3"},{"comment":"The paper reports at p = 0.020 that the transition temperature keeps decreasing with lattice volume and that the infinite-volume limit of T_c may not exist, yet all p_c values are quoted from finite lattices up to 24^3 without finite-size scaling or an infinite-volume extrapolation. If T_c can fail to have an infinite-volume limit at one nonzero p, the same possibility applies to the other quoted thresholds, and the central estimates could shift substantially. The authors should provide a finite-size analysis for each threshold or explicitly state that the quoted values are finite-lattice estimates with an assessment of their uncertainty.","section":"Section 2, paragraph on p = 0.020; Section 4"},{"comment":"The central mapping and all threshold values inherit the noise model in Eqs. (1)-(3), but the derivation of these error probabilities is deferred to Ref. [11], and the key assumption that all qubit errors share a single probability p and that CNOT error propagation reduces to the quoted anisotropic rates is stated without justification here. Because the physical meaning of the thresholds depends entirely on this simplification, the manuscript needs at least a concise derivation of Eqs. (1)-(3) or an explicit validation against circuit-level simulations; otherwise the reader cannot assess whether the mapped gauge theory describes the toric/surface code under realistic noise.","section":"Eqs. (1)-(3)"}],"minor_comments":[{"comment":"There are typographical errors: \"Mont Carlo\" should be \"Monte Carlo\" in the abstract and introduction, and \"sufrace code\" in the Section 3 heading should be \"surface code\".","section":"Abstract and Section 1"},{"comment":"The caption contains \"the the wrong-sign probability\" and should be corrected.","section":"Fig. 5 caption"},{"comment":"The text says the Nishimori line is the red dotted line in Fig. 5, but the figure as printed does not appear to show any red dotted line; please clarify the figure or the text.","section":"Fig. 5"},{"comment":"The symbol pr(I) is used in the Nishimori condition but is not defined; it should be defined as the probability of no error on a qubit.","section":"Eq. (4)"},{"comment":"The lattice sizes in Fig. 5 are typeset as \"8 3\", \"12 3\", etc.; these should be rendered as 8^3, 12^3, and so on for readability.","section":"Fig. 5 legend"}],"recommendation":"major_revision","confidential_remarks":"This is clearly a proceedings contribution that depends on the companion paper [11] for the derivation of the noise model and mapping. The editor should confirm that [11] is available or in press, because the present manuscript does not stand alone on that point. The inconsistent descriptions of the plotted noise model and the absence of finite-size analysis are the main obstacles; if the authors can align the figures with the text and add a proper extrapolation or an explicit disclaimer, the paper could be considered further."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a proceedings write-up of the companion paper [11], where the RCPGM and the full threshold investigation actually live. The three quoted thresholds (0.00682, 0.06, 0.0144) are attributed to [11], not derived here. What this paper adds is a Monte Carlo study using the Polyakov line as the gauge-invariant order parameter, plus susceptibility and third-cumulant curves on lattices up to 24^3. That is a legitimate and sensible tool for this problem, and the prose is honestly labeled preliminary. The explicit admission that at p=0.020 the transition temperature keeps decreasing with volume and may have no infinite-volume limit is the kind of caveat you want to see.\n\nThe soft spots are real and, in one place, serious. The paper reports no statistical uncertainties and no infinite-volume extrapolation for any of the p_c values, so the numbers are best treated as finite-lattice estimates. The stress-test concern is accurate: Section 2, around Figs. 2-4, describes a well-above-threshold case at p=0.00852 and a near-threshold case at p=0.00682 for bit-flip plus syndrome noise, but the actual figure captions and Section 3 state that Figs. 2-4 are circuit-level noise with p=2.88e-5 and p=2.31e-2, with threshold p_c ~ 0.0144. No plot at p=0.00682 or p=0.00852 is shown. So the threshold 0.00682 is not tied to any displayed data, and the reader cannot check the central claim from this paper alone. The thresholds 0.06 and 0.0144 are quoted without supporting curves. The mapping and the error-probability formulas are deferred to [11], which is fine as a citation strategy but makes this paper a pointer rather than a self-contained result.\n\nI do not see circularity: the p_c values are outputs, not fitted inputs, and the Nishimori condition is standard. The self-citation to [10] and [11] is appropriate because those are the papers containing the model and the details. Minor typos (\"Mont Carlo\", \"sufrace\", \"Y-erorr\") confirm the proceedings-level polish.\n\nWho gets value from this? A lattice-field-theory reader wanting a quick impression of how Polyakov-line observables work for this mapping, or someone deciding whether to read [11]. It is not a reliable source for the threshold values themselves.\n\nFor peer review: I would send it to a referee for the proceedings, conditionally. The internal inconsistency between Section 2 and the figure captions must be fixed, and the p_c quotes should either come with error bars and a finite-size discussion or be explicitly deferred to [11]. A referee should also verify that the companion paper supplies the missing numerical support.","headline":"A candid proceedings summary of the companion RCPGM paper [11], with honest caveats but internal inconsistencies that sever the quoted thresholds from the displayed data.","tokens_in":8677,"tokens_out":1952,"would_cite":false,"duration_ms":19755,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P70","82B20","82B80"],"pacs":["03.67.Pp","11.15.Ha"],"model":"deepseek-v4-flash","headline":"This paper maps the toric/surface code's decoding problem under realistic circuit-level noise to Z(2) lattice gauge theories, and finds via Monte Carlo simulation that the threshold probabilities are p_c ≈ 0.00682, 0.06, and 0.0144 for…","keywords":["quantum error correction","toric code","surface code","lattice gauge theory","Monte Carlo simulation","threshold probability","depolarizing noise","Z(2) gauge theory"],"falsifier":"Run the same Monte Carlo simulation on a lattice larger than $24^{3}$ at p=0.020 and at the claimed threshold values, tracking the Polyakov-line susceptibility and third-order cumulant; if the transition temperature continues to decrease with volume and the zero-crossing of the third-order cumulant disappears, the threshold estimates are finite-size artifacts rather than genuine phase boundaries. A complementary check is a direct circuit-level simulation of the toric/surface code under the same noise models: if a near-optimal decoder cannot reproduce logical error rates that drop only below these p_c values, the mapping is not quantitatively correct.","tokens_in":7632,"feed_emoji":"🛡️","tokens_out":11053,"duration_ms":86653,"temperature":0.7,"pith_summary":"The paper argues that the decoding problem for Kitaev's toric/surface code under realistic circuit-level noise can be mapped onto three-dimensional Z(2) lattice gauge theories, so the question of whether fault-tolerant quantum error correction is possible becomes a question about the phase diagram of those theories. Monte Carlo simulation of the mapped models, run on the Nishimori line where the quenched disorder matches the thermal ensemble, yields threshold error probabilities p_c ≈ 0.00682 for bit/phase-flip noise plus syndrome noise, p_c ≈ 0.06 for uniform depolarizing noise plus syndrome noise, and p_c ≈ 0.0144 for anisotropic depolarizing circuit-level noise plus syndrome noise. These values sit above the thresholds obtained with Minimum-Weight Perfect Matching decoding, which the paper takes as evidence that better decoders could improve the practical fault tolerance of the toric/surface code. The order parameters used are gauge-invariant Polyakov-line observables, required because the mapped models are Z(2) gauge theories.","feed_headline":"Toric-code thresholds beat MWPM in new simulation","feed_subtitle":"Z(2)-gauge mapping sets surface-code thresholds at 0.0068, 0.06, 0.0144—above MWPM, so better decoders may help.","key_machinery":"The load-bearing object is the random coupled-plaquette gauge model (RCPGM): a three-dimensional $Z(2)\\times Z(2)$ lattice gauge theory with Hamiltonian $H = \\sum_n [H_X(n)+H_Y(n)+H_Z(n)]$, where Ising spins $\\sigma$ and $\\tau$ live on space-time links and plaquette couplings $J$ are assigned wrong signs according to the error probabilities. Qubit bit-flip/phase-flip errors become wrong-sign spatial plaquette couplings, while syndrome measurement errors become wrong-sign temporal couplings. The Nishimori line, $\\exp(-4|J(W)|) = \\mathrm{pr}(X)\\mathrm{pr}(Y)\\mathrm{pr}(Z)/(\\mathrm{pr}(W)^2\\,\\mathrm{pr}(I))$ and $\\exp(-2|J_{t\\sigma,\\tau}|) = q/(1-q)$, fixes the disorder distribution to the thermal distribution, and the threshold is read off where this line meets the phase boundary. Because these are gauge theories, local order parameters are forbidden by Elitzur's theorem, so the paper uses the Polyakov line $P(i,j)=\\prod_t \\sigma_t(i,j)$, a gauge-invariant product of Ising spins along the time direction, and locates the thermal transition by its susceptibility and third-order cumulant; parallel tempering is used to equilibrate the frustrated quenched-disorder systems.","core_discovery":"The central claim is that the random coupled-plaquette gauge model (RCPGM), a three-dimensional Z(2)×Z(2) lattice gauge theory with anisotropic and random-sign couplings, captures the main aspects of toric/surface code error correction under depolarizing and syndrome noise, and that its phase transition marks the quantum error correction threshold. On the Nishimori line, the phase boundary between the ordered phase (where decoding succeeds) and the disordered phase (where it fails) terminates at the threshold probability. The Monte Carlo results place these thresholds at p_c ≈ 0.00682 (X/Z noise plus syndrome), p_c ≈ 0.06 (isotropic depolarizing plus syndrome), and p_c ≈ 0.0144 (anisotropic circuit-level depolarizing plus syndrome). Because these estimates are higher than the corresponding MWPM thresholds, the paper concludes that current practical decoding algorithms are not optimal and leave room for improvement.","pith_inferences":["If the mapping is quantitatively accurate, then current surface-code experiments reporting 'below threshold' operation may be limited by classical decoding algorithms rather than by quantum hardware; a decoder that approximates the optimal statistical-mechanics decoder could raise the operating noise rate substantially.","The paper's observation that T_c at p=0.020 decreases with lattice volume may indicate that some noise models have no finite threshold in the thermodynamic limit, which would impose a hard ceiling on error rates irrespective of code size; the paper reports the observation but still quotes p_c values without an infinite-volume extrapolation.","Noise symmetrization may be a practical lever: if depolarizing noise can be manufactured by twirling or randomized compiling, the far higher isotropic threshold (p_c ≈ 0.06) suggests a route to cheaper fault tolerance than engineering anisotropic noise away.","A direct check of the mapping's quantitative predictions would be to run small toric codes under the same three noise models with a decoder derived from the gauge-theory phase boundary, and compare the observed logical error rates to the predicted threshold location."],"forward_implications":["If the paper is right, the toric/surface code's practical threshold is set as much by the decoder as by the code: MWPM leaves a gap that improved decoders could close.","The large difference between the isotropic threshold (p_c ≈ 0.06) and the anisotropic circuit-level threshold (p_c ≈ 0.0144) shows that the structure of the noise, not just its total rate, determines whether fault-tolerant operation is possible.","Adding syndrome measurement noise forces the statistical model to be three-dimensional, so error correction thresholds should be viewed as properties of the space-time error history rather than of a static code layout.","The same statistical-mechanics mapping and Monte Carlo method can be applied to other topological codes, such as color codes, which the paper names as the next target."],"supporting_citations":[{"why":"Shows how topological quantum memory can be mapped to a statistical mechanics model, the foundation for identifying the ordered phase with successful decoding.","marker":"[9]"},{"why":"Prior work by the same collaboration deriving fundamental thresholds of realistic QEC circuits from classical spin models, which this proceedings extends.","marker":"[10]"},{"why":"Defines the random coupled-plaquette gauge model and supplies the error-probability assignments (Eqs. 1–3) that the Monte Carlo simulation uses.","marker":"[11]"},{"why":"Gives the Nishimori condition that equates disorder and thermal distributions, used to set the line along which thresholds are extracted.","marker":"[12]"},{"why":"Provides the parallel tempering algorithm needed to equilibrate the frustrated quenched-disorder spin systems.","marker":"[13]"},{"why":"Elitzur's theorem, invoked to justify using gauge-invariant observables (the Polyakov line) rather than local order parameters.","marker":"[14]"},{"why":"Introduces the disordered gauge theory and confinement-Higgs-type transition analysis of the accuracy threshold, a direct ancestor of the phase-diagram approach used here.","marker":"[15]"},{"why":"Documents finite-size scaling at first-order transitions, used to interpret the volume dependence of the transition temperature near the threshold.","marker":"[20]"}],"fun_headline_variants":["Toric code thresholds beat MWPM via Z(2) gauge mapping","Surface-code thresholds exceed MWPM in Z(2) gauge simulation","Z(2) lattice gauge model yields higher error thresholds than MWPM","Monte Carlo Z(2) simulation raises toric-code thresholds above MWPM","Better decoders on horizon: surface-code thresholds beat MWPM"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the simplified error probabilities in Eqs. 1–3 faithfully represent how circuit-level Pauli errors propagate through the CNOT gates of the syndrome measurement circuit; the quoted thresholds also assume that finite-lattice transition temperatures converge to a thermodynamic limit, even though the paper reports that at p=0.020 the transition temperature keeps decreasing with lattice volume.","fun_headline_variants_meta":{"raw":{"variants":["Toric code thresholds beat MWPM via Z(2) gauge mapping","Surface-code thresholds exceed MWPM in Z(2) gauge simulation","Z(2) lattice gauge model yields higher error thresholds than MWPM","Monte Carlo Z(2) simulation raises toric-code thresholds above MWPM","Better decoders on horizon: surface-code thresholds beat MWPM"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000248,"raw_usage":{"total_tokens":1545,"prompt_tokens":945,"completion_tokens":600,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":561,"completion_tokens_details":{"reasoning_tokens":505}},"tokens_in":561,"tokens_out":600,"duration_ms":5471,"temperature":1.0,"reasoning_tokens":505,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T15:46:56.746366+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same Monte Carlo simulation on a lattice larger than $24^{3}$ at p=0.020 and at the claimed threshold values, tracking the Polyakov-line susceptibility and third-order cumulant; if the transition temperature continues to decrease with volume and the zero-crossing of the third-order cumulant disappears, the threshold estimates are finite-size artifacts rather than genuine phase boundaries. A complementary check is a direct circuit-level simulation of the toric/surface code under the same noise models: if a near-optimal decoder cannot reproduce logical error rates that drop only below these p_c values, the mapping is not quantitatively correct.","supporting_citations":[{"cited_title":"Exact results on spin glass models","cited_arxiv_id":null,"evidence_quote":"Gives the Nishimori condition that equates disorder and thermal distributions, used to set the line along which thresholds are extracted."},{"cited_title":"Parallel tempering: Theory, applications, and new perspectives","cited_arxiv_id":null,"evidence_quote":"Provides the parallel tempering algorithm needed to equilibrate the frustrated quenched-disorder spin systems."},{"cited_title":"Impossibility of spontaneously breaking local symmetries","cited_arxiv_id":null,"evidence_quote":"Elitzur's theorem, invoked to justify using gauge-invariant observables (the Polyakov line) rather than local order parameters."},{"cited_title":"Confinement-Higgs transition in a disordered gauge theory and the accuracy threshold for quantum memory","cited_arxiv_id":null,"evidence_quote":"Introduces the disordered gauge theory and confinement-Higgs-type transition analysis of the accuracy threshold, a direct ancestor of the phase-diagram approach used here."},{"cited_title":"Finite-sizescalingatfirst-orderphasetransitions","cited_arxiv_id":null,"evidence_quote":"Documents finite-size scaling at first-order transitions, used to interpret the volume dependence of the transition temperature near the threshold."}],"review_version":1}