{"id":"c65099e3-81dc-4fb4-bfca-e601ebe714f7","arxiv_id":"2507.19313","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The authors present the complete leading-order one-loop amplitude for gg to HHH in closed analytic form, with full top and bottom quark mass dependence, validated against two independent numerical programs.","lead":"This paper derives the first fully analytic one-loop amplitude for triple Higgs production via gluon fusion, including full dependence on the mass of the quark in the loop. The result is compact and fast, enabling direct probes of the Higgs self-couplings and a practical building block for future higher-order calculations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reconstruction completeness rests on the unproved conjecture that the fourth primary component of ideal (2.19) has no compact generators; the box remainders in eqs. (3.35) and (3.40) are exposed to this gap.","rationale":"The central claim is that the analytic expressions in Sections 2, 3, and 4 exactly reproduce the one-loop gg→HHH amplitude with full mass dependence. The authors support this with cross-checks against OpenLoops and Recola2 and by releasing Fortran and Python implementations. Those are real and valuable pieces of evidence. The reader's conditional verdict is therefore reasonable: the result is very likely correct in practice. The weakest point is not the physics of the amplitude itself but the completeness of the analytic reconstruction. The paper explicitly admits that one primary component of ideal (2.19) is only conjectured to be primary and that no compact generators are known. That ideal is directly connected to the Δ_{12×3×4} denominators appearing in two of the most complicated box remainders. If the unknown component contains additional algebraic structure, the finite-field ansatz could be missing a piece of the coefficient functions. The provided test script does not cover this fourth component. This is a formal gap in the derivation, not an observed error. It does not warrant rejection, because the numerical agreement and the reproducibility of the code give strong practical assurance. A secondary concern is the sign discrepancy with Ref. [1] in the 14 TeV parametrization; this is worth documenting but is less load-bearing for the central analytic amplitude, since it likely reflects a typo in the paper's phenomenological parametrization or in the reference. The verdict should remain CONDITIONAL, with the primary-component conjecture and the 14 TeV sign discrepancy as the two items to be clarified.","tokens_in":19652,"tokens_out":6869,"duration_ms":71353,"concrete_test":"Use the ideal-quotient construction (which the authors state is available) to compute explicit generators of the fourth primary component of the ideal in eq. (2.19). Then, at a dense set of finite-field points on that component's variety, test whether the numerators of the box remainders d^{-+}_{12×3×4} and d^{++}_{12×3×4} vanish modulo the component. If any numerator does not vanish, the ansatz is missing a required factor and the published coefficients are incomplete. As a complementary numerical check, compare the analytic matrix element against OpenLoops at 10^4 random and near-singular phase-space points using quadruple precision; a mismatch would confirm the gap, though a match would not formally close it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing gap is the unproved primary-decomposition conjecture in Section 2. The reconstruction in eqs. (2.15)-(2.16) claims the integral coefficients are fully determined by finite-field and p-adic evaluations. This claim depends on knowing all denominator and numerator structures that can appear. For the ideal <⟨1|5|4|3|2], Δ_{12×3×4}> in eq. (2.19), the paper identifies three simple components and then states that the fourth component is 'conjecture[d] to be primary' and that no compact generators are available. The box remainders that carry Δ_{12×3×4} denominators, namely eqs. (3.35) and (3.40), are precisely the coefficients most exposed to this gap. If the unknown component contains additional factors on which the true numerator does not vanish, the fitted ansatz could omit a term or a denominator, and the 'fully analytic' claim would be incomplete at some phase-space points even though the OpenLoops and Recola2 spot checks pass. The ancillary test script verifies eqs. (2.18) and (2.20), not the fourth component, so the gap is explicitly acknowledged but not closed. This does not mean the result is wrong; the numerical cross-checks are genuine evidence. However, for a formal proof of completeness, this is the weakest link.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a compact analytic result for the leading-order one-loop amplitude for gg -> HHH, retaining full dependence on the mass of the heavy quark in the loop. The amplitude is decomposed into triangle, box, and pentagon contributions, each reduced to scalar integrals with coefficients obtained by Passarino-Veltman reduction and simplified by finite-field and p-adic analytic reconstruction, partial fraction decomposition, and primary decomposition. The central result is the set of coefficient formulas in Sections 2 and 3, together with a Fortran implementation and a Python-readable version of the coefficients. The authors report full agreement with OpenLoops and Recola2 for the matrix element, an order-of-magnitude speed improvement, and cross-section parametrizations at 14 and 100 TeV. The paper also provides a script checking two of the three displayed primary decompositions used in the reconstruction.","tokens_in":19886,"tokens_out":6426,"duration_ms":61375,"significance":"If correct, this is a significant technical achievement: it provides the first compact analytic representation of a 2 -> 3 one-loop amplitude with three massive external scalars and full heavy-quark mass dependence, relevant for probes of the triple and quartic Higgs self-couplings. The result is supported by genuinely independent numerical cross-checks against two public one-loop codes, and the ancillary code and validation scripts are valuable reproducibility artifacts. The claimed speed and stability of the analytic expression could be important for future NLO computations and for phenomenological studies of triple Higgs production. The main caveat is a formally unproved completeness conjecture in the primary decomposition used by the reconstruction; this does not invalidate the numerical cross-checks, but it does bear directly on the 'fully analytic' claim and should be addressed explicitly in a revision.","major_comments":[{"comment":"The completeness of the coefficient reconstruction rests on the conjecture, stated immediately after eq. (2.20), that the remaining primary component of the ideal in eq. (2.19) is primary and admits no compact generating set. The box remainders in eqs. (3.35) and (3.40) carry explicit factors of Delta_12x3x4 and are precisely the terms exposed to this gap: if the true numerator has support on the unknown component in a way not captured by the finite-field and p-adic evaluations, the ansatz could miss a denominator or a term at isolated phase-space points. The attached script test_primary_decompositions.py checks eqs. (2.18) and (2.20) but not the fourth component, so this gap is acknowledged but not closed. I ask the authors to either prove the needed completeness statement or, failing that, to state the conditional nature of the 'fully analytic' claim explicitly and to add targeted numerical checks of eqs. (3.35) and (3.40) at phase-space points lying on the variety of the unknown component.","section":"Section 2, eqs. (2.19)-(2.20)"},{"comment":"The 14 TeV cross-section parametrization in eq. (4.3) differs from the result quoted in Ref. [1] in the sign of the Delta_kappa_4 term, while the 100 TeV parametrization in eq. (4.4) agrees with Ref. [34]. A sign difference in a small coefficient can change the qualitative interpretation of quartic-coupling sensitivity, so this is not a purely typographical matter. Please verify the cross-section evaluation, identify the source of the difference (input conventions, PDFs, scale choices, or an error), and correct or explicitly discuss the discrepancy in the revised manuscript.","section":"Section 4, eqs. (4.3)-(4.4)"},{"comment":"The abstract and Section 4 claim that evaluation of the new amplitude is 'an order of magnitude or more faster' than OpenLoops and Recola2, but no benchmark protocol is provided. Please report the hardware, compiler, number of phase-space points, whether the comparison is for a single phase-space evaluation or for an integrated cross section, and the numerical precision achieved. The speed and stability claims are part of the advertised value of the result and should be reproducible.","section":"Section 4, after eq. (4.1)"}],"minor_comments":[{"comment":"The sign convention for tr5 as Tr{p1 p2 p3 p4 gamma5} should be stated explicitly, since the transformation rule in eq. (3.21) flips its sign and subsequent formulas depend on this convention.","section":"Section 2, eq. (2.14)"},{"comment":"The notation '|M^2|' is an abuse of notation; please write |\\mathcal{M}|^2 and define \\mathcal{M} once, for clarity.","section":"Section 4, eq. (4.1)"},{"comment":"The trace notation 'tr(5|4|3|1-2)' is introduced in eq. (3.41) just before its use in eq. (3.40); consider defining it before the first occurrence and use 'cf.' instead of 'c.f.'.","section":"Section 3.3, eq. (3.41)"},{"comment":"The mapping from the permutations in eqs. (2.26)-(2.28) to the coefficient labels in eq. (2.25) is implicit; one sentence stating that each coefficient inherits the momentum ordering of the corresponding integral would make the sums unambiguous.","section":"Section 2.3, eqs. (2.25)-(2.28)"}],"recommendation":"major_revision","confidential_remarks":"The central amplitude is likely correct: the independent agreement with OpenLoops and Recola2 is strong evidence, and the attached code increases confidence. My main reservation is formal: the unproved primary-decomposition conjecture in Section 2 leaves the completeness claim of the reconstruction technically open, and the box remainders (3.35) and (3.40) are the natural place where an omitted factor could hide. I believe this can be addressed by either a proof, a carefully stated completeness assumption with additional targeted checks, or a softened claim. The 14 TeV sign discrepancy with the HHH whitepaper should also be resolved before publication. The paper is well within the scope of JHEP."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper delivers the first fully analytic, all-mass-dependent one-loop amplitude for gg→HHH, and the central result is solid. The reconstruction completeness has a formally acknowledged gap, but the numerical cross-checks make me confident the formulas are right.\n\nThe genuinely new thing is the analytic form itself: compact expressions for the triangle, box, and pentagon contributions with full top- and bottom-mass dependence, evaluated an order of magnitude faster than OpenLoops or Recola2. The authors cross-check against those two independent codes and find full agreement, and they ship a Fortran implementation plus a Python version in a Zenodo repo. That is reproducible evidence, and it earns real credit.\n\nThe soft spots are two, and neither undermines the main result. First, Section 2's reconstruction relies on a conjecture: the fourth primary component of the ideal <⟨1|5|4|3|2], Δ_{12×3×4}> is stated to be primary but no compact generators are given. The box remainders in (3.35) and (3.40) carry Δ_{12×3×4} denominators and are exactly the terms exposed to that gap. If the missing component contained extra numerator factors, the ansatz could be incomplete at some phase-space points. The authors acknowledge this openly, and the numerical checks pass, so I read this as a limitation of the proof, not a sign the result is wrong. But the phrase \"fully analytic\" is a bit stronger than what is formally established. Second, the 14 TeV parametrization differs from the HHH whitepaper by the sign of the Δκ4 term; the 100 TeV parametrization agrees. The authors note this but don't explain it. That is a small red flag in the phenomenology section and should be resolved before publication.\n\nWho is this for? Anyone working on triple-Higgs phenomenology or on analytic one-loop amplitudes with massive particles. The techniques are an incremental extension of the authors' previous reconstruction program, but the specific result is new and useful. I would cite it.\n\nMy recommendation: send it to peer review. The two issues are clarifications, not showstoppers. A good referee should push for a comment on the conjecture's practical impact and for a word on the sign discrepancy.","headline":"A genuinely useful analytic one-loop amplitude for gg->HHH, well cross-checked, with an acknowledged but minor completeness gap in the reconstruction and a small unresolved sign discrepancy.","tokens_in":20424,"tokens_out":2096,"would_cite":true,"duration_ms":19072,"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":"The paper derives the full one-loop amplitude for triple-Higgs production in gluon fusion in closed analytic form, with exact dependence on the loop-quark mass.","keywords":["triple Higgs production","gluon fusion","one-loop amplitude","analytic reconstruction","Higgs self-couplings","spinor-helicity formalism","pentagon reduction","finite-field methods"],"falsifier":"Evaluate the closed-form amplitude against an independent direct numerical evaluation of the one-loop Feynman diagrams at an ensemble of random phase-space points in high precision, especially near $\\mathrm{tr}_5\\to0$ and $\\Delta_{12\\times3\\times4}\\to0$; any mismatch beyond integration accuracy would show that the reconstruction missed a term.","tokens_in":19393,"feed_emoji":"⚛️","tokens_out":10292,"duration_ms":98960,"temperature":0.7,"pith_summary":"The paper establishes that the leading-order one-loop amplitude for triple Higgs production in gluon fusion, $gg\\to HHH$, can be written in closed analytic form with exact dependence on the mass of the quark circulating in the loop. The full amplitude is organized into triangle, box, and pentagon sub-amplitudes, and the pentagon contribution, the genuinely new part, is reduced to compact coefficients built from spinor brackets and the parity-odd trace $\\mathrm{tr}_5$. This matters because $gg\\to HHH$ is a direct probe of the triple and quartic Higgs self-couplings, and a fast closed form replaces repeated numerical loop integrations. The paper reports exact numerical agreement with independent fully numerical one-loop programs and a speed gain of more than an order of magnitude.","feed_headline":"Closed-form gg→HHH one-loop amplitude derived","feed_subtitle":"Exact quark-mass dependence plus order-of-magnitude speed: the ingredient NLO triple-Higgs predictions need.","key_machinery":"The load-bearing device is analytic reconstruction in a covariant polynomial quotient ring (eqs. 2.15 and 2.16), in which the on-shell Higgs momenta are treated as variables modulo the mass-shell relations; finite-field and $p$-adic evaluations at carefully chosen slices determine the coefficient functions. The simplification step uses the Cayley-matrix pentagon reduction, with coefficients $c^{(i)}=-\\frac12\\sum_j S^{-1}_{ij}$, plus a rewriting that inserts a factor of unity built from the relation between the Cayley and Gram determinants, converting pentagon coefficients into 'effective pentagon' coefficients in which powers of $\\mathrm{tr}_5^2$ cancel explicitly. Primary decompositions of ideals of spinor-string expressions, such as those in eqs. (2.17) and (2.19), expose common numerator factors, and the resulting compact coefficients are assembled with a library of scalar one-loop integrals.","core_discovery":"The paper's central claim is that the complete one-loop QCD amplitude for $0\\to ggHHH$, written as $A_{\\mathrm{tot}}=\\delta^{AB}\\frac{g_s^2}{16\\pi^2}\\frac{m^4}{v^3}(A_3+A_4+A_5)$, is reproduced exactly by the analytic formulas of Sections 2 and 3: explicit spinor-helicity expressions for the triangle coefficients (eqs. 2.21 and 2.24), the box coefficients (eqs. 2.23 and 2.24), the effective pentagon coefficients (eqs. 3.19, 3.20, 3.22, 3.23), and the box and triangle remainder coefficients (eqs. 3.31 through 3.49). The formulas keep full dependence on the loop-quark mass $m$ and the Higgs mass $M_H$, and the resulting matrix element reproduces the one-loop result at any phase-space point, with cross-checks against fully numerical one-loop programs. Evaluated at $M_H=125$ GeV with top and bottom quarks, the Standard Model cross sections are $0.0512$ fb at 14 TeV and $2.76$ fb at 100 TeV, and the dependence on the Higgs self-coupling deviations $\\Delta\\kappa_3$, $\\Delta\\kappa_4$ is compressed into the polynomial parametrizations of eqs. (4.3) and (4.4).","pith_inferences":["Editorial extension: expanding the analytic coefficients in inverse powers of $m^2$ would give the exact heavy-top-limit corrections order by order, providing a direct test of how much the $m\\to\\infty$ approximation shifts kinematic distributions.","Editorial extension: the same reduction-and-reconstruction machinery is directly applicable to the real-radiation processes $0\\to q\\bar q gHHH$ and $0\\to gggHHH$, so the method likely supplies all the pieces needed for an exact NLO calculation of triple-Higgs production.","Editorial extension: the quartic-coupling sensitivity visible in the parametrizations is mild (coefficient of $\\Delta\\kappa_4$ near $-0.1$), suggesting that extracting $\\kappa_4$ from inclusive $gg\\to HHH$ will require differential measurements or higher energy, both of which become cheaper with an analytic amplitude.","Editorial extension: at phase-space boundaries where $\\mathrm{tr}_5$ or the Gram determinants vanish, the explicit cancellations in these formulas could be validated in high precision; if they hold, the amplitude should remain stable in double-precision Monte Carlo without numerical rescue."],"forward_implications":["The exact leading-order matrix element can serve directly as the Born term in a future NLO calculation of triple-Higgs production, replacing the heavy-top-limit reweighting that the paper argues distorts mass effects in double-Higgs production.","The Standard Model cross sections at this order, $0.0512$ fb at 14 TeV and $2.76$ fb at 100 TeV, together with the polynomial fits in eqs. (4.3) and (4.4), allow experimental groups to convert limits on $\\kappa_3$ and $\\kappa_4$ into cross-section constraints without recomputing the loop amplitude.","Because full quark-mass dependence is retained, approximate NLO predictions can be reweighted with the exact mass dependence rather than with $m\\to\\infty$ effective-theory factors.","The compact formula evaluates more than an order of magnitude faster than fully numerical one-loop programs, making it practical for event generation and large Monte Carlo samples."],"supporting_citations":[{"why":"Supplies the systematic method for extracting analytic one-loop amplitudes from numerical evaluations, which is the foundation of the reconstruction workflow.","marker":"[51]"},{"why":"Provides the ansatz-construction machinery over $p$-adic numbers and algebraic geometry used to fix the coefficient functions.","marker":"[52]"},{"why":"Earlier analytic one-loop amplitudes for Higgs plus three partons; supplies the ideal decomposition used for canceling denominators.","marker":"[54]"},{"why":"Defines the 'scalar-tops' covariant reconstruction setup that this paper adapts to three equal-mass on-shell Higgs states.","marker":"[55]"},{"why":"Passarino-Veltman reduction provides the initial decomposition of each diagram into scalar integrals.","marker":"[56]"},{"why":"Gives the closed pentagon-to-box reduction coefficients via the inverse Cayley matrix, used in eq. (3.5).","marker":"[58]"},{"why":"Introduces the strategy of rewriting pentagon coefficients with an inserted factor of unity so Gram-determinant factors cancel, leading to the 'effective pentagon' construction.","marker":"[59]"},{"why":"Scalar one-loop integral library used to turn the analytic coefficients into numerical matrix elements.","marker":"[60]"},{"why":"Fully numerical one-loop implementation used as an independent cross-check and as the baseline for the speed comparison.","marker":"[61]"},{"why":"Second independent fully numerical implementation used as a cross-check and speed comparison.","marker":"[62]"}],"fun_headline_variants":["Analytic one-loop gg→HHH amplitude","Closed-form triple Higgs with full mass dependence","Exact gg→HHH amplitude: 10x faster evaluation","Full-mass analytic amplitude for triple Higgs production","One-loop gg→HHH: exact, compact, fast"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The formulas are guaranteed only if every coefficient is a rational function of the listed spinor and invariant variables; the paper itself notes it could not find compact generators for one algebraic component arising in the reconstruction, so an undocumented square-root or algebraic factor there would make the expressions incomplete.","fun_headline_variants_meta":{"raw":{"variants":["Analytic one-loop gg→HHH amplitude","Closed-form triple Higgs with full mass dependence","Exact gg→HHH amplitude: 10x faster evaluation","Full-mass analytic amplitude for triple Higgs production","One-loop gg→HHH: exact, compact, fast"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00026,"raw_usage":{"total_tokens":1606,"prompt_tokens":981,"completion_tokens":625,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":597,"completion_tokens_details":{"reasoning_tokens":549}},"tokens_in":597,"tokens_out":625,"duration_ms":5989,"temperature":1.0,"reasoning_tokens":549,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:55:55.476204+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the closed-form amplitude against an independent direct numerical evaluation of the one-loop Feynman diagrams at an ensemble of random phase-space points in high precision, especially near $\\mathrm{tr}_5\\to0$ and $\\Delta_{12\\times3\\times4}\\to0$; any mismatch beyond integration accuracy would show that the reconstruction missed a term.","supporting_citations":[{"cited_title":"Passarino and M.J.G","cited_arxiv_id":null,"evidence_quote":"Passarino-Veltman reduction provides the initial decomposition of each diagram into scalar integrals."}],"review_version":1}