{"id":"cc2f0916-f44c-4858-a416-fd0fec1bfa98","arxiv_id":"1908.10950","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A constrained HMC algorithm with the Rattle integrator measures the derivative of the effective Higgs potential over the full field range in 4D Abelian-Higgs and 5D SU(2) gauge theories.","lead":"The paper develops a constrained Hybrid Monte Carlo algorithm, based on the Rattle integrator, that fixes the average Higgs field during lattice simulations of gauge-Higgs models. It allows the effective Higgs potential to be measured over its full range with a precision that does not degrade with volume.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"5D gauge-link Rattle updates lack an explicit reversibility check, leaving the gauge-field central claim unverified.","rationale":"I read the paper as a methods contribution whose central claim is Eq. (2.5) plus the demonstration that constrained HMC with Rattle makes this observable precise over the full domain. The 4D Abelian-Higgs evidence is strong: the h->0 limit of the lambda observable is shown in Fig. 2, the potential matches the one-loop result when m_H is fitted, and Table 2 shows volume-independent errors. The load-bearing risk is therefore the generalization to gauge fields. If the 5D link updates are not reversible (or not measure-preserving), the Metropolis step does not converge to the constrained Gibbs measure, and the measured <lambda> is not the derivative of the constraint effective potential; this would invalidate the paper's headline claim to have solved the problem for gauge theories. The reader's weakest_assumption identifies exactly this gap. I found no internal inconsistency in the 4D derivation or numerics; the formal issue in the short derivation of Eq. (2.3) is a presentation shorthand, since the final relation is the standard constraint-force identity and is numerically validated. The missing 5D reversibility test is the single check that would settle whether the gauge-field extension actually works, so the CONDITIONAL verdict is appropriate and unchanged.","tokens_in":24936,"tokens_out":13058,"duration_ms":147813,"concrete_test":"Run the same time-reversal test as Sec. 3.1 on the 5D torus Rattle update at beta4=beta5=1.66, Omega=8^4, N5=4, Phi=0.4: integrate one trajectory with +h, then continue with -h, and require return to the initial (U,pi) to machine precision; repeat for h=0.05 and h=0.01. If the map fails this test, the 5D results in Fig. 6 do not sample the constrained ensemble and the gauge-field claim collapses; if it passes, the evidentiary gap is closed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim extends the constrained-HMC/Lagrange-multiplier method to theories with gauge fields. That extension rests on the Rattle updates for SU(2) links derived in Secs. 4, 5 and Appendices A.2-A.4. In Sec. 3.1 the authors numerically check time-reversibility and volume preservation for the 4D Abelian-Higgs case, but no such check is reported for the 5D torus or orbifold updates. The 5D updates determine lambda^(1) by a Secant root-find on a transcendental constraint (Eq. 4.2c, A.2.2c); if multiple roots exist or the selected root depends on integration direction, the map is not reversed by flipping h, and the Metropolis acceptance no longer samples the constrained path integral used in Eq. (2.5). The statement in Sec. 6 that the algorithm was numerically tested for time-reversibility does not specify that the tests cover the gauge-link cases. This is an evidentiary gap, not a demonstrated failure; the 4D results and volume-scaling Table 2 stand independently.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops constrained Hybrid Monte Carlo algorithms for gauge-Higgs models using an extension of the Rattle integrator to non-linear and gauge-field constraints. The central theoretical result is Eq. (2.5), which expresses the derivative of the constraint effective potential as the expectation value of the first Lagrange multiplier: U'_Ω(Φ) = -⟨λ^(1)⟩_Φ/Ω. The method is implemented for the 4D Abelian-Higgs model, where reversibility and det J = 1 are tested numerically, and effective potentials are compared against histogram results and a one-loop perturbative potential. The same framework is then adapted to 5D SU(2) gauge theory on the torus (with constraints on ⟨TrP⟩ and ⟨(TrP)^2⟩) and on the orbifold (constraint on ⟨TrP⟩), with first results for the constraint effective potentials. The paper claims, for the first time, a solution to constrained HMC for theories with gauge fields.","tokens_in":25201,"tokens_out":8216,"duration_ms":79948,"significance":"If the claims hold, the method is a notable technical advance: it allows precise determination of the constraint effective potential over the full range of the Higgs/Polyakov-loop variable, with statistical errors that do not grow with volume, unlike histogram methods. The derivation of Eq. (2.5) is clean and self-contained, and the 4D numerical tests (reversibility, det J = 1, h→0 limit, agreement with correlator masses) are convincing. The 5D extension to non-canonical gauge-link updates is novel and potentially important for gauge-Higgs unification and effective Polyakov-loop actions. However, the 5D results currently lack explicit reversibility and volume-preservation tests, and the one-loop comparison is partly self-fulfilling because m_H is fit. These issues are local and fixable rather than fatal.","major_comments":[{"comment":"The 5D Rattle updates for SU(2) gauge links are asserted to preserve time-reversibility and volume preservation by citing the general Rattle theory of Refs. [24,25]. The numerical tests reported in Sec. 3.1, however, cover only the 4D Abelian-Higgs model, whose updates are canonical. The 5D updates are non-canonical because the link update is multiplicative (U_{n+1} = exp(h π_{n+1/2}) U_n) and λ^(1) is solved by a Secant root-find on a transcendental equation (Sec. 4.2c and A.2.2c). The unqualified statement in Sec. 6 that the algorithms were 'numerically tested for time-reversibility and volume preservation' is therefore not supported for the 5D cases. Please add explicit reversibility and det J tests for the 5D torus (both constraints) and orbifold updates, or supply a proof that the Rattle geometric properties extend to the Lie-group update used here.","section":"Secs. 4, 5 and Appendices A.2-A.4"},{"comment":"The statement that the constraint effective potential agrees well with the one-loop Higgs potential is weakened by the fact that m_H is a fit parameter: the derivative U'_1loop(Φ) is fitted to the measured U'_Ω(Φ), so the agreement is partly self-fulfilling. The independent correlator masses m_H,R in Table 1 do provide support (e.g., m_H = 1.093(1) vs m_H,R = 1.099(21) for β=8, κ=0.166), but the paper does not quantify the fit quality or state the number of fit parameters and the Φ-range used. Please report χ²/dof, the fit range, and explicitly compare the fitted m_H with m_H,R for each parameter set.","section":"Sec. 3.3, Eq. (3.8), Table 1"},{"comment":"The choice of the minus sign in front of the square root in Eq. (3.6c) is justified only by the statement that 'during numerical simulations it turns out that only the − sign fulfills the constraint condition.' Because the Rattle map must be single-valued for reversibility, the branch ambiguity should be resolved analytically; the plus sign can in fact be excluded by the requirement of a finite h→0 limit, as follows from the expansion leading to Eq. (3.7). For the 5D Secant root-finding (Eq. 4.2c and A.2.2c), please comment on the uniqueness of the root and demonstrate that the selected root is the physical one, e.g., by a reversibility check on representative configurations.","section":"Eq. (3.6c) and Sec. 4.2c"}],"minor_comments":[{"comment":"The orbifold results use the constraint Φ = ⟨TrP/2⟩, which the authors themselves state is not the exact Higgs field H = (1/4Ω)∑Tr[P5−P5†,σ3]^2 (Sec. 5). Please say this explicitly in the abstract or conclusions, and avoid the phrase 'full domain of the Higgs variable' for the orbifold case.","section":"Abstract and Sec. 5"},{"comment":"The text after Eq. (3.11) contains a typo: 'wich' should be 'which'.","section":"Sec. 3.3"},{"comment":"The affiliation contains a typo: 'Comuter Science' should be 'Computer Science'.","section":"Title page"},{"comment":"The caption contains a typo: 'compcat phase' should be 'compact phase'.","section":"Fig. 7 caption"},{"comment":"The derivation of the initial guess for the Secant method is shown to O(h^2); please state the order of the truncation and report how the Secant iteration converges (typical number of iterations, achieved machine precision).","section":"Eq. (4.3) and A.2.4"},{"comment":"The 4^4 row reports Φ0 = 1.122(8), which differs from the larger-volume values 1.132-1.133(1); the text says 'except for the smallest volume we don't see an effect.' Please clarify whether this is a finite-volume effect or a statistical fluctuation, and whether the quoted mH fit for 4^4 uses the same Φ-range as the other volumes.","section":"Table 2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a serious extension of the authors' proceedings [28], with a clean central derivation and convincing 4D tests. The main gap is the missing numerical verification of reversibility and volume preservation for the 5D gauge-link Rattle updates; this is an evidentiary gap rather than a demonstrated failure, and it can be fixed with additional tests or a proof. The one-loop comparison should be reframed as a one-parameter consistency check. The paper is within the scope of the journal and, after revision, could be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth reading. This is the first constrained-HMC implementation for gauge fields, and the core observable, the derivative of the constrained effective potential as minus the expectation value of the first Lagrange multiplier, is derived cleanly in Eqs. (2.3)-(2.5). The 4D Abelian-Higgs implementation is the strongest part: the authors check time-reversibility, verify det J = 1, show the h -> 0 limit of the multiplier observable, and demonstrate volume-independent precision against the histogram method in Table 2. That alone is a solid methods result.\n\nThe genuinely new content is the adaptation of Rattle to non-canonical gauge-link dynamics with composite Polyakov-loop constraints, for both the 5D torus and the orbifold. That is a real extension, and the 5D potentials shown are plausible. The one-loop comparison in unitary gauge is useful, though the fit of m_H to the measured derivative is partly self-fulfilling; the extracted mass does agree with the correlator mass m_H,R in Table 1, so there is independent support.\n\nThe soft spot is the 5D section. Section 6 says time-reversibility was numerically tested, but the explicit checks shown are for the 4D case only. For the 5D updates, lambda^(1) is determined by a Secant root-find on a transcendental constraint. If the selected root depends on integration direction or multiple roots exist, the map may not invert properly, and the Rattle literature guarantees for standard Hamiltonian systems do not automatically transfer to the non-canonical gauge-link form. This is an evidentiary gap, not a demonstrated failure; the 4D results and volume-scaling table stand independently. A minor related point: the orbifold constraint is a proxy for the true Higgs field, and the authors acknowledge this.\n\nCitation pattern is fair. Kuti-Shen and Hairer-Lubich-Wanner are properly credited, and the novelty claim is accurate: the combination of the multiplier observable with Rattle on gauge links is not in the cited literature.\n\nThis paper deserves a serious referee. I would recommend conditional acceptance with a request for an explicit numerical reversibility and volume-preservation check on the 5D updates, or at least a precise statement of which tests were run. The central claim is likely correct; the support just needs to reach the 5D cases.","headline":"Solid methods paper: the Lagrange-multiplier observable is clean and the 4D implementation is verified, but the 5D gauge-link extension lacks the explicit reversibility check that would make the central claim airtight.","tokens_in":25682,"tokens_out":1952,"would_cite":true,"duration_ms":21812,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Measuring the constraint force itself, rather than a histogram of field values, resolves the full effective Higgs potential on the lattice at volume-independent precision.","keywords":["constrained hybrid Monte Carlo","Rattle algorithm","constraint effective potential","gauge-Higgs models","Abelian-Higgs model","five-dimensional SU(2) gauge theory","Polyakov loop","lattice gauge theory"],"falsifier":"Run a single Rattle trajectory for the 5D torus model at $\\beta_4 = \\beta_5 = 1.66$ with the Polyakov-loop constraint, then reverse it step by step with stepsize $-h$; if the SU(2) link variables do not return to their initial values up to round-off, the constrained update is not reversible and the measured derivative $-\\langle\\lambda^{(1)}\\rangle/\\Omega$ would not be the derivative of a well-defined potential.","tokens_in":24733,"feed_emoji":"","tokens_out":8243,"duration_ms":77530,"temperature":0.7,"pith_summary":"The paper develops constrained Hybrid Monte Carlo algorithms that keep a chosen Higgs observable fixed during the simulation, and shows that the derivative of the resulting constraint effective potential is directly given by the expectation value of the Lagrange multiplier enforcing the constraint. This turns potential measurement from a histogram of field fluctuations into a local observable, so the potential can be computed over the entire allowed range of the Higgs variable instead of only a narrow region around its unconstrained expectation value. The precision of the new observable does not deteriorate when the lattice volume is increased, which is what makes infinite-volume and mass determinations feasible. The method is implemented for the 4D Abelian-Higgs model and for 5D SU(2) gauge theory on a torus and on an orbifold, and in the Abelian-Higgs case the measured potential agrees with the one-loop continuum Higgs potential in unitary gauge. A sympathetic reader should take the paper's central claim to be that this observable is a practical, quantitative route to effective potentials in gauge-Higgs theories.","feed_headline":"One Lagrange multiplier maps the full Higgs potential","feed_subtitle":"A constrained HMC algorithm computes the effective potential over the full Higgs domain, with errors that do not grow with volume","key_machinery":"The load-bearing object is the Rattle algorithm, an extension of the leap-frog/St\\\"ormer-Verlet method that adds a half step for the momenta so that fields and momenta are available at the same integration time; this makes it possible to impose the hidden constraint and solve for the two Lagrange multipliers $\\lambda^{(1)}$ and $\\lambda^{(2)}$. For the 4D Abelian-Higgs model the constraint is the spatial average of $\\phi^\\dagger \\phi$, and $\\lambda^{(1)}$ can be written in closed form (taking the minus branch); for the 5D SU(2) torus and orbifold cases the constraint is built from Polyakov loops, and $\\lambda^{(1)}$ is found by a Secant iteration seeded by a Taylor expansion of the exponential link update. The observable $U'_\\Omega(\\Phi) = -\\langle\\lambda^{(1)}\\rangle_\\Phi/\\Omega$ carries the actual physics: $\\lambda^{(1)}$ is the force that holds the Higgs observable fixed, and its average is the derivative of the potential. Auxiliary machinery includes orthogonal projection of the initial random momenta onto the hidden-constraint surface and a continuum-form rewriting of $\\lambda^{(1)}$ that removes the dependence on the unphysical integration step size.","core_discovery":"Starting from the constrained path integral, the paper derives $U'_\\Omega(\\Phi) = -\\frac{1}{\\Omega} \\langle \\lambda^{(1)} \\rangle_\\Phi$, where $\\lambda^{(1)}$ is the Lagrange multiplier that holds the averaged Higgs field at value $\\Phi$. The central discovery is that this identity can be used as an observable: during a constrained HMC simulation one measures the multiplier and obtains the derivative of the constraint effective potential directly, with no histogram binning and no need to fit a distribution. To make the identity usable for gauge theories, the authors extend the Rattle algorithm to constrained Hamiltonian systems whose equations of motion are non-canonical for SU(2) gauge links, solving for the multipliers at each molecular-dynamics step (first $\\lambda^{(1)}$ from the constraint, then $\\lambda^{(2)}$ from the hidden constraint). They demonstrate that the continuum limit in integration step size exists and that the resulting potentials reproduce the histogram result where the latter is available, while extending it over the full domain of the Higgs field and keeping errors roughly constant as the volume grows. They also verify the 4D result against the one-loop Higgs potential in unitary gauge, extracting Higgs masses consistent with two-point function determinations.","pith_inferences":["Because the derivative is a local expectation value rather than a histogram slope, one could in principle join measurements from simulations at different $\\Phi$ values into a continuous potential with controlled interpolation error; the paper does not itself perform such a reconstruction.","The reversibility argument for the 5D non-canonical gauge-link updates is inherited from the general Rattle theory; an explicit numerical reversibility test on the torus or orbifold would settle whether the hidden constraint is preserved exactly there, and would be a natural companion check.","A direct extension suggested by the method is to constrain other order parameters, such as topological charge or the density of a conserved current, and read off the corresponding effective potential from the same Lagrange-multiplier identity.","If the volume-independence of the precision persists on larger lattices, the main cost of the method shifts to the Secant solve for the multiplier, making algorithmic tuning of that solve (or a closed-form approximation) the practical bottleneck; the paper does not analyze this scaling."],"forward_implications":["In the 4D Abelian-Higgs model, the derivative of the constraint effective potential can be measured over the full range of the Higgs variable, including regions where the potential diverges ($\\Phi \\to 0$), whereas histograms only cover a neighbourhood of the expectation value.","The statistical error of the derivative observable stays approximately constant as the lattice volume grows, so volume extrapolation to the infinite-volume effective potential becomes practical; the Higgs mass can then be read off from the curvature at the minimum.","For the 5D SU(2) torus model the method reproduces the Mexican-hat form of the potential for the Polyakov-loop Higgs field and confirms the two degenerate minima of $\\langle \\operatorname{Tr} P \\rangle$, which the histogram method cannot resolve.","For the 5D orbifold model, the constrained potential can be measured across the whole interval $-1 < \\operatorname{Tr} P / 2 < 1$, giving access to the confining, Higgs, and hybrid phases from one observable.","The same constrained algorithms can be applied to effective Polyakov loop actions, where individual Polyakov lines are constrained locally rather than globally, providing a route to finite-temperature and finite-density QCD effective potentials."],"supporting_citations":[{"why":"Defines the constraint effective potential and its infinite-volume limit, the quantity the paper measures.","marker":"[19]"},{"why":"Establishes that the constraint potential converges to the conventional effective potential, justifying the lattice measurement.","marker":"[20]"},{"why":"Introduced measuring the derivative of the constraint effective potential through the Lagrange multiplier, the core observable used here.","marker":"[21]"},{"why":"Supplies the Hybrid Monte Carlo algorithm that the constrained updates are built on.","marker":"[22]"},{"why":"Provides the Rattle algorithm and its reversibility, symplecticity, and volume-preservation properties that the constrained integrator relies on.","marker":"[24]"},{"why":"Supplies the geometric integration theory for St\\\"ormer-Verlet and Rattle methods that justifies the time-reversible constrained flow.","marker":"[25]"},{"why":"Gives the one-loop Abelian-Higgs potential in unitary gauge used to verify the measured constraint potential.","marker":"[23]"},{"why":"Provides the variable transformation to unitary gauge that lets the comparison with the one-loop potential be made.","marker":"[26]"}],"fun_headline_variants":["Full Higgs potential from one Lagrange multiplier","Constraint HMC maps whole Higgs domain","No histograms: full effective potential via HMC","Lagrange multiplier gives full Higgs potential","Full-domain effective Higgs potential with constant errors"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the constrained molecular-dynamics update being exactly time-reversible in the five-dimensional gauge cases, but the paper demonstrates reversibility numerically only for the four-dimensional Abelian-Higgs model and relies on the general Rattle theory for the SU(2) updates.","fun_headline_variants_meta":{"raw":{"variants":["Full Higgs potential from one Lagrange multiplier","Constraint HMC maps whole Higgs domain","No histograms: full effective potential via HMC","Lagrange multiplier gives full Higgs potential","Full-domain effective Higgs potential with constant errors"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000189,"raw_usage":{"total_tokens":1370,"prompt_tokens":1014,"completion_tokens":356,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":630,"completion_tokens_details":{"reasoning_tokens":291}},"tokens_in":630,"tokens_out":356,"duration_ms":4508,"temperature":1.0,"reasoning_tokens":291,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:36:10.352041+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a single Rattle trajectory for the 5D torus model at $\\beta_4 = \\beta_5 = 1.66$ with the Polyakov-loop constraint, then reverse it step by step with stepsize $-h$; if the SU(2) link variables do not return to their initial values up to round-off, the constrained update is not reversible and the measured derivative $-\\langle\\lambda^{(1)}\\rangle/\\Omega$ would not be the derivative of a well-defined potential.","supporting_citations":[{"cited_title":"Derivation of the Effective Potential,","cited_arxiv_id":null,"evidence_quote":"Defines the constraint effective potential and its infinite-volume limit, the quantity the paper measures."},{"cited_title":"The Constraint Effective Potential,","cited_arxiv_id":null,"evidence_quote":"Establishes that the constraint potential converges to the conventional effective potential, justifying the lattice measurement."},{"cited_title":"Supercomputing the Effective Action,","cited_arxiv_id":null,"evidence_quote":"Introduced measuring the derivative of the constraint effective potential through the Lagrange multiplier, the core observable used here."},{"cited_title":"Hybrid Monte Carlo,","cited_arxiv_id":null,"evidence_quote":"Supplies the Hybrid Monte Carlo algorithm that the constrained updates are built on."},{"cited_title":"Geometric numerical integration illustrated by the St¨ ormer-Verlet method,","cited_arxiv_id":null,"evidence_quote":"Supplies the geometric integration theory for St\\\"ormer-Verlet and Rattle methods that justifies the time-reversible constrained flow."},{"cited_title":"Montvay and G","cited_arxiv_id":null,"evidence_quote":"Provides the variable transformation to unitary gauge that lets the comparison with the one-loop potential be made."}],"review_version":1}