{"id":"d332592d-2606-41fe-a86c-ddf34d2956e1","arxiv_id":"2606.27835","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Defect density under weak symmetry breaking is the standard Kibble-Zurek power law times an exponential suppression set by the probability of rare fluctuations that drive a domain into the disfavored state.","lead":"This paper explains why forming defects during a phase transition becomes exponentially harder when a small external bias weakly breaks the symmetry. It derives a formula for the defect density using rare-event probability and confirms it in simulations of one- and two-dimensional models.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Core derivation of P_wrong is a scaling assertion; the promised Freidlin-Wentzell check is omitted, leaving the central exponent vulnerable.","rationale":"The reader's weakest-assumption identifies the heuristic derivation of P_wrong in Eq. (3), and I agree that this is the single most load-bearing step: the entire exponential correction in Eq. (5) rests on the assertion that the optimal rare fluctuation is a coherent noise of magnitude h over the full KZ spacetime volume. The paper explicitly promises but omits the Freidlin-Wentzell minimization, so the exponents nu(d+z)/(1+z nu) are not derived. A quick linearized analysis indicates the optimal noise profile is time-dependent and concentrated near the critical point, not a constant over Omega_KZ; while the integrated action may still scale as Eq. (4) because the variance of the domain average is reduced by xi_hat^{-d}, that cancellation is nontrivial and must be shown. The simulations in Fig. 2 provide supporting evidence, but they are not a substitute for the derivation, especially because the assumed mean-field critical exponents in d=1,2 are not obviously the true exponents of the stochastic GL model. I do not see an internal contradiction that proves Eq. (5) wrong; the concern is a serious gap in proof, not a demonstrated error. Therefore the reader's CONDITIONAL verdict is appropriate, and no change is needed. The proposed FW boundary-value calculation would either close the gap or overturn the central claim.","tokens_in":1044,"tokens_out":2234,"duration_ms":273237,"concrete_test":"For the linearized stochastic Ginzburg-Landau equation in d=1 with b(t)=t/tau_Q, numerically solve the Euler-Lagrange boundary-value problem for the minimum of the Onsager-Machlup action S = integral zeta^2/(4 eta theta) dx dt subject to the equation of motion and the terminal constraint that the average of phi over a KZ domain of size xi_hat = eta^{-1/4} tau_Q^{1/4} at t_hat = eta^{1/2} tau_Q^{1/2} lies on the disfavored side. Extract the scaling S ~ (h^2/theta) tau_Q^{alpha} eta^{beta} over a range of tau_Q, eta, theta, h and compare alpha with nu(d+z)/(1+z nu) = (d+2)/4 = 3/4 for nu=1/2, z=2. If alpha != 3/4, Eq. (5) is falsified; if alpha = 3/4, the heuristic is confirmed despite the omitted FW derivation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is Eq. (5), whose exponential action S_h is derived in Eqs. (3)-(4) by assuming that the rare event is a coherent thermal fluctuation zeta~h occupying the entire KZ spacetime volume Omega_KZ = t_hat xi_hat^d. This is a heuristic scaling identification, not a calculation. The true large-deviation rate is the minimum of the Onsager-Machlup/Freidlin-Wentzell action over all noise histories that drive a KZ domain into the disfavored state. For the linearized dynamics eta d_t phi = (t/tau_Q) phi + h + zeta, the optimal noise is not constant: it is proportional to the adjoint Green's function e^{((t_hat^2 - t^2)/(2 eta tau_Q))}, peaked near the critical point t=0 with width ~t_hat. The action's scaling with tau_Q, eta, and h is not obviously the same as that of a constant force over Omega_KZ; it survives only because the variance of the domain-averaged field is suppressed by xi_hat^{-d}, which is exactly the calculation the paper omits. The manuscript states that a formal FW minimization 'yields the same result' but does not show it. If the optimal profile has an effective duration or volume different from t_hat xi_hat^d, the exponent nu(d+z)/(1+z nu) in Eq. (5) changes and the central claim fails. This is the keystone of the paper, and its derivation is absent.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper addresses defect formation in continuous phase transitions when a weak explicit symmetry-breaking field h is present, where recent numerics show an exponential correction to Kibble–Zurek scaling. The authors propose that this correction is a rare-event problem: with probability P_wrong, a Kibble–Zurek domain of spacetime volume Ω_KZ = t̂ξ̂^d is driven by a coherent thermal fluctuation ζ∼h into the disfavored symmetry-broken state. This leads to the action S_h ∼ (h^2/θ)(τ_Q/η)^{ν(d+z)/(1+zν)} and the closed-form prediction Eq. (5) for the defect density in arbitrary dimensions. The paper reports simulations of stochastic Ginzburg–Landau models in 1D and 2D which show an exponential suppression of defect density and a collapse of the fitted suppression coefficient λ onto the scaling variable h^2/(θη^{(d+2)/4}).","tokens_in":5069,"tokens_out":4685,"duration_ms":53488,"significance":"If the central prediction Eq. (5) is correct, the paper provides a missing general framework for a recently observed numerical phenomenon and makes a falsifiable, universality-class-dependent prediction. The master-curve collapse in Figs. 2(d,e) is genuinely encouraging, and the combination of a closed-form formula with direct simulation is a strength. However, the derivation of the exponential action S_h is not actually carried out: the optimal-fluctuation profile is assumed rather than derived, and the promised Freidlin–Wentzell minimization is only stated. In addition, the numerical verification fixes the τ_Q exponent before extracting λ, so the confirmation of the central exponent is weaker than it appears. The significance is therefore conditional on the missing large-deviation calculation being supplied and on a more independent numerical test of the τ_Q scaling.","major_comments":[{"comment":"The central step of the paper is the identification of the exponential action S_h. The sentence 'A domain can end in the disfavored state only if a coherent thermal fluctuation of magnitude ζ∼h persists throughout the entire freeze-out spacetime volume' is a scaling assumption, not a derivation. The subsequent statement that a formal Freidlin–Wentzell minimization 'yields the same result' is not shown. This matters because the optimal noise history for the linearized dynamics η∂_tφ = (t/τ_Q)φ + h + ζ is not obviously a constant over Ω_KZ; it is proportional to the adjoint Green's function and is peaked near t=0. If the effective duration or spatial profile of the optimal fluctuation differs from t̂ξ̂^d, the exponent ν(d+z)/(1+zν) in Eq. (5) would change. The authors should either provide the Freidlin–Wentzell calculation, including the optimal noise profile and the resulting action, or g","section":"Eqs. (3)–(4), text below Eq. (2)"},{"comment":"The numerical verification does not independently test the τ_Q exponent in Eq. (5). The suppression coefficient λ is extracted by fitting the simulation data to n_defects = n_KZ exp[−λ τ_Q^{(d+2)/4}], so the exponent (d+2)/4 is already imposed before λ is obtained. The collapse in Figs. 2(d,e) then confirms the h-, θ-, and η-dependence of λ, but not the τ_Q dependence, once the nonuniversal constant A_d is adjusted. The reported 'S_h ∝ τ_Q^{3/4}' in Fig. 2(c) should be presented as a fit with error bars and, ideally, with λ extracted from a fit that does not fix the exponent a priori. Without this, the claim that simulations verify the predicted exponential scaling is overstated.","section":"Figs. 2(c)–2(e) and text near 'We find that the action cost...'"}],"minor_comments":[{"comment":"Typo: 'vacuam' should be 'vacua'.","section":"Introduction, first paragraph"},{"comment":"The path-probability functional P[ζ] is written without the normalization factor. This is harmless for the scaling argument, but including the partition function would make the large-deviation statement more precise.","section":"Eq. (2)"},{"comment":"The label 'Action cost S_h' is used for what appears to be numerically extracted −log(n_defects/n_KZ) or a related quantity. Please clarify the definition used in the plot, especially because the y-axis values are not dimensionless and no error bars are shown.","section":"Fig. 2(c)"},{"comment":"The paper says Eq. (5) holds 'in arbitrary dimensions,' but the derivation is heuristic and the simulations are limited to d=1,2. A d=3 test would strengthen the claim, or the claim should be softened to 'any d in the same universality class.'","section":"Eq. (5) and the paragraph before it"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses an interesting and timely problem, and the proposed rare-event picture is plausible. My main concern is not novelty but rigor: the central action S_h is assumed, not derived, and the omitted Freidlin–Wentzell calculation is exactly what would justify the exponent in Eq. (5). The numerical collapse is suggestive but partly rests on fixing the exponent under test. I would like to see the large-deviation derivation and a more independent numerical test of the τ_Q scaling before publication. If the authors cannot supply the FW calculation, the manuscript would remain a plausible hypothesis rather than a demonstrated theory."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: if Eq. (5) is right, this is a real advance. It turns a puzzling numerical suppression into a closed-form scaling law. The mechanism — rare coherent noise pushing a KZ domain into the disfavored vacuum — is new relative to Ref. [3], which only reported the numerics. I suspect the formula is correct, but the paper needs to show the calculation it says it has up its sleeve.\n\nThe paper does two things well. It derives a concrete exponential correction to KZ, with an exponent set by the universal ν and z, and it tests that prediction in 1D and 2D simulations. The collapse of the extracted suppression coefficient onto the predicted scaling variable is a genuine check of the h, θ, η dependencies, not just a one-parameter fit. The authors are also upfront that A_d is non-universal and fitted, and that the derivation is heuristic.\n\nWhere I worry: the central probability P_wrong in Eq. (3) is an assertion that the optimal fluctuation is a constant force h over the whole KZ spacetime volume. That is plausible for a linearized problem, but it is not a derivation. The authors say a Freidlin–Wentzell minimization gives the same result and then don't show it. That omission matters because the exponential's power in τ_Q is the paper's key quantitative prediction. The stress-test note suggests the optimal noise is peaked near the critical point rather than constant; my own back-of-the-envelope for the linearized dynamics suggests the action scaling may survive because the relevant variance is the domain average, but the paper owes the reader that calculation. Without it, the central result is a scaling estimate, not a demonstrated theorem.\n\nMinor but real: no error bars, no simulation details, and the use of mean-field exponents (ν=1/2, z=2) in d=1,2 is not justified. The models are Ginzburg–Landau with thermal noise, and in low dimensions the equilibrium exponents are not mean-field. The h=0 KZ scaling seems to match the simulations, so this is probably a Gaussian fixed point effect, but it should be said. Also, A_d is fitted, so the test only constrains the scaling exponents; that is acceptable, but the paper shouldn't oversell absolute prediction.\n\nThe circularity concern in the reader's report is milder than it looks. The τ_Q exponent is verified independently in Fig. 2(c), and the collapse tests the other dependences. The free constant is non-universal, so nothing inappropriate there.\n\nWho this is for: people doing quenches in systems with weak fields, and anyone working on large-deviation effects at phase transitions. It deserves a serious referee. I'd send it to review, but insist that the FW calculation be added, or at least a proper derivation of the optimal noise profile.\n\nNet: worth engaging, needs a revision to be fully convincing.","headline":"A plausible rare-event extension of Kibble-Zurek for weak symmetry breaking, with a clean scaling prediction and supporting numerics, but the keystone action is asserted rather than derived.","tokens_in":5469,"tokens_out":6833,"would_cite":true,"duration_ms":66181,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Defect formation under weak symmetry breaking is controlled by rare thermal fluctuations that bias entire Kibble–Zurek domains, yielding an exponential correction to the standard quench scaling.","keywords":["rare events","large deviations","Kibble-Zurek mechanism","defect formation","symmetry breaking","stochastic Ginzburg-Landau","quench dynamics","topological defects"],"falsifier":"Compute the optimal noise history by numerically minimizing the large-deviation action for a stochastic Ginzburg–Landau quench and check whether the minimizer is indeed a constant noise of magnitude ~h over the Kibble–Zurek volume. If the optimal path is non-uniform, the scaling in Eq. (5) is falsified. Equivalently, measure the probability P_wrong that a domain selects the disfavored state in a 1D quench and test whether log P_wrong ∝ (h^2/θ)(τ_Q/η)^{3/4} across a broad range of parameters.","tokens_in":4589,"feed_emoji":"🎲","tokens_out":4487,"duration_ms":52603,"temperature":0.7,"pith_summary":"The paper claims that when a continuous phase transition is crossed in the presence of a weak symmetry-breaking field, defect formation is no longer governed by random domain choices but by rare thermal fluctuations that drive entire correlated domains into the disfavored symmetry-broken state. This turns the problem into a large-deviation problem, producing a closed-form expression for the defect density in any dimension: the usual Kibble–Zurek power law multiplied by an exponential suppression factor. The suppression exponent depends only on the field strength, noise strength, damping, and quench time through a universal combination of critical exponents. The authors verify the prediction numerically in one- and two-dimensional stochastic Ginzburg–Landau models, finding collapse onto the predicted scaling. If correct, this unifies a previously unexplained numerical correction to Kibble–Zurek scaling and provides a general framework for defect formation in realistic systems where symmetry is never perfect.","feed_headline":"Rare events, not random choice, set defect counts in biased quenches","feed_subtitle":"A weak symmetry-breaking field exponentially suppresses defect formation via a single rare-noise action, verified in 1D and 2D.","key_machinery":"The central object is the Kibble–Zurek domain: a correlated region of size ξ̂ whose symmetry-breaking choice is frozen at the quench timescale. The argument then computes the probability that such a domain ends up in the disfavored state by estimating the cost of a rare thermal fluctuation in Gaussian noise. That cost is the action S_h = (h^2/(4ηθ)) Ω_KZ, where Ω_KZ ~ t̂ ξ̂^d is the freeze-out spacetime volume. Using the critical scalings t̂ ~ τ_Q^{zν/(1+zν)} and ξ̂ ~ τ_Q^{ν/(1+zν)} converts this into the closed-form exponent in Eq. (5). The mechanism is explicitly distinguished from nucleation theory: no activation over a free-energy barrier is involved; instead, a whole domain is coherentl","core_discovery":"The central claim is that under weak explicit symmetry breaking, the probability for a Kibble–Zurek domain to select the disfavored vacuum is exponentially small and set by a rare-event action S_h ~ (h^2/θ)(τ_Q/η)^{ν(d+z)/(1+zν)}. The defect density then becomes n_defects ~ (η/τ_Q)^{dν/(1+zν)} exp[−A_d (h^2/θ)(τ_Q/η)^{ν(d+z)/(1+zν)}]. This expression reduces to standard Kibble–Zurek scaling when the field h is zero, and in the weak-field, weak-noise limit it predicts an exponential suppression that cannot be captured by any perturbative correction. The derivation assumes that the rare event is a coherent thermal noise of magnitude ~h persisting throughout the entire freeze-out spacetime volu","pith_inferences":["If the rare-event picture is correct, the same exponential suppression should appear in non-mean-field universality classes (e.g., Wilson–Fisher) but with the exponent ν(d+z)/(1+zν) evaluated with the appropriate ν and z, a prediction that could be tested in classical or quantum quench experiments.","The action S_h may be measurable directly from the probability distribution of domain choices in Monte Carlo simulations, giving a direct probe of the large-deviation functional without needing to count rare defects.","The result suggests practical ways to control topological defect densities in ultracold atomic gases or superconducting systems by applying a weak external field, exponentially suppressing unwanted defects even for slow quenches.","The assumption of uniform coherent noise can be relaxed; if the optimal fluctuation profile is non-uniform, the exponent in Eq. (5) would change, offering a route to test the theory's core mechanism through path-integral minimization."],"forward_implications":["Defect densities in quenched systems with weak symmetry breaking should show an exponential falloff with quench time, not just a power law, with an exponent set by the combination ν(d+z)/(1+zν).","The correction factor is non-universal in its prefactor A_d but universal in its scaling combination, so the scaling collapse of the suppression coefficient λ versus h^2/(θ η^{(d+2)/4}) should persist across different microscopic models in the same universality class.","The framework extends straightforwardly to defects of arbitrary dimensionality, so the same exponential suppression is expected for strings and membranes, not just point defects.","In the limit of very weak field or very fast quench, the exponential factor approaches one and standard Kibble–Zurek scaling is recovered, providing a smooth crossover between the two regimes.","The distinction from nucleation corrections means that the exponential suppression here appears even when the metastable minimum is only slightly disfavored and no barrier crossing is required."],"fun_headline_variants":["Rare noise events set defect counts in biased quenches","Weak field exponentially cuts defects via rare fluctuation","Rare-event action sets exponential defect suppression","Defect density suppressed by rare noise in biased quenches","Rare fluctuations decide defect counts under weak bias"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper assumes that the rare event that flips a domain is a single coherent thermal fluctuation of size ~h acting uniformly over the entire freeze-out space-time volume; if the optimal fluctuation has nontrivial spatial or temporal structure, the predicted exponential exponent would no longer hold.","fun_headline_variants_meta":{"raw":{"variants":["Rare noise events set defect counts in biased quenches","Weak field exponentially cuts defects via rare fluctuation","Rare-event action sets exponential defect suppression","Defect density suppressed by rare noise in biased quenches","Rare fluctuations decide defect counts under weak bias"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000487,"raw_usage":{"total_tokens":2219,"prompt_tokens":711,"completion_tokens":1508,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":455,"completion_tokens_details":{"reasoning_tokens":1433}},"tokens_in":455,"tokens_out":1508,"duration_ms":10666,"temperature":1.0,"reasoning_tokens":1433,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T09:52:18.934297+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the optimal noise history by numerically minimizing the large-deviation action for a stochastic Ginzburg–Landau quench and check whether the minimizer is indeed a constant noise of magnitude ~h over the Kibble–Zurek volume. If the optimal path is non-uniform, the scaling in Eq. (5) is falsified. Equivalently, measure the probability P_wrong that a domain selects the disfavored state in a 1D quench and test whether log P_wrong ∝ (h^2/θ)(τ_Q/η)^{3/4} across a broad range of parameters.","supporting_citations":[],"review_version":2}