{"id":"a02c3e84-3bff-4afc-a5d8-769772e76a73","arxiv_id":"2502.03282","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Two-loop helicity amplitudes for diphoton production with a massive top quark loop are computed, providing new analytic results for gluon fusion and benchmark numbers for both channels.","lead":"This paper computes two-loop QCD helicity amplitudes for diphoton production while keeping the full top-quark mass in the loop, and reports analytic results for the gluon-fusion channel. These amplitudes are needed for more precise diphoton cross-section predictions at the high-luminosity LHC.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"IBP reduction completeness is the load-bearing assumption: the unified system initially missed three master-integral relations, so a further missed mapping would invalidate the coefficients; the only finite-value cross-check is an unpublished placeholder.","rationale":"The paper is a careful technical calculation with strong internal checks, but the load-bearing step is indeed the completeness of the IBP reduction. The authors' own report that three relations had to be added by hand is the clearest indicator that the automated reduction is not guaranteed to be exhaustive. The finite remainders are the new physical result, and they are not directly checked except against a yet-unpublished reference with a placeholder number. The UV/IR cancellation and Bose symmetry checks are necessary but not sufficient to validate the finite coefficients. The correct verdict remains CONDITIONAL: the calculation is credible and well structured, but the reduction completeness should be verified with an independent solver, and the independent comparison needs to become publicly inspectable before the claim can be fully accepted.","tokens_in":16906,"tokens_out":12123,"duration_ms":124821,"concrete_test":"Run an independent IBP reduction with FIRE6 (or LiteRed) on all uncrossed families and their five crossings for the gg channel, using the same denominator definitions as Table 1, and verify that (i) the per-family master-integral counts are 29/32/54/36, (ii) the three hand-added relations from Section 3 are generated automatically, and (iii) the reduction of a representative sample of top-sector form factors reproduces the ancillary-file coefficients to 10 significant digits. If FIRE6 finds additional relations beyond the three, or if any coefficient differs, the reduction is incomplete and the amplitudes would need revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result is a linear combination of master integrals obtained by IBP reduction of 26,577 (gg) and 2,289 (qq-bar) scalar integrals over four families and five crossings. Section 3 admits that this reduction is not fully automatic: the unified system initially returned 91 master integrals and the authors had to add three further relations by hand, stating that modern IBP software overlooks some mappings between integrals. The correctness of every coefficient in the ancillary file therefore depends on the authors having found all such missed relations. The checks in Section 5 do not close this gap: UV/IR pole cancellation and Bose symmetry are sensitive to the divergent parts and to discrete symmetries, but not to the finite rational coefficients that the reduction determines. The only finite-value check is against ref. [92], whose arXiv identifier is a placeholder and whose results cannot currently be inspected. If a fourth relation was missed, the master-integral set would be overcomplete in a way that changes the coefficients, and the claimed first analytic two-loop amplitudes would be incorrect without being caught by any check in the paper.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript computes two-loop QCD helicity amplitudes for gg→γγ and q qbar→γγ with full top-quark mass dependence in the loop. The amplitudes are decomposed via physical projectors, reduced to master integrals using IBP relations, and expressed analytically in terms of polylogarithmic and elliptic functions. The paper presents the UV renormalization (on-shell for the top mass, MS for other quantities), IR factorization (applied only to the gluon channel), benchmark finite remainders at one phase-space point per channel, and bare amplitudes in an ancillary file. The authors claim the first analytic two-loop helicity amplitudes for the gluon-fusion channel with a massive quark loop.","tokens_in":17033,"tokens_out":18292,"duration_ms":177435,"significance":"If correct, the result is a valuable building block for diphoton production at NNLO/N3LO with top-quark mass effects and has direct applications to dijet production. The paper has several strengths: a transparent helicity-projector setup, explicit handling of the elliptic master integrals, a consistent renormalization scheme, and multiple internal consistency checks (UV cancellation, IR pole matching, Bose symmetry). The bare amplitudes are also provided in an ancillary file, which supports reproducibility. The main weaknesses are the reliance on a manually completed IBP reduction that initially missed relations, the unavailability of the independent numerical cross-check, and the limited set of public benchmark finite remainders.","major_comments":[{"comment":"The manuscript admits that the unified IBP system initially returned 91 master integrals and that three additional relations had to be supplied by hand because modern IBP software overlooks some mappings between integrals. The checks reported in Section 5 are insensitive to a missed finite rational coefficient in the master-integral expansion: UV/IR pole cancellation and Bose symmetry probe only divergent parts and discrete symmetries, not the finite coefficients of the master integrals. A further missed relation would therefore alter every coefficient in the ancillary file while passing all shown checks. Please document the three added relations explicitly and provide an independent validation of the reduction, for example by numerically evaluating both sides of the reduced expressions at several random phase-space points using AMFlow or by comparing against a second independent reduction.","section":"Section 3, IBP reduction paragraph"},{"comment":"The claim of perfect numerical agreement with ref. [92] is the only finite-value external check in the paper, but ref. [92] is cited with the placeholder arXiv number 2501.xxxx and its results cannot currently be inspected. Please update the reference if the work has appeared, include a table with the comparison values, or otherwise make the cross-check verifiable. Without this, the finite remainders are supported only by a non-public agreement.","section":"Section 5, first paragraph"},{"comment":"The paper states without derivation that q qbar → γγ does not exhibit any IR divergences and therefore applies IR subtraction only to the gluon channel. Since the external quarks are massless and on-shell, the absence of IR poles is not self-evident; a short argument or a reference to the analogous statement in ref. [25] is needed to rule out soft and collinear singularities in the H^f and H^ft contributions. The finite remainders in Table 2 depend directly on this point.","section":"Section 4.2, eqs. (4.12)-(4.18)"}],"minor_comments":[{"comment":"The captions give 's = 3 GeV and N = 3'. The units of s are ambiguous; the introduction says the benchmarks are around the top-quark threshold, which s = 3 GeV is not. Please specify whether s is in units of m_t^2 (for example, s = 3 m_t^2) and define N (presumably N_c).","section":"Tables 2 and 3"},{"comment":"Reference [92] is listed with the placeholder identifier 2501.xxxx; the placeholder should be removed and replaced by a complete citation before publication.","section":"References"},{"comment":"The finite remainder is said to be 'available upon request from the authors.' Given the paper's claims, attaching the finite remainders in electronic form, or at least the numerical evaluation code, would improve reproducibility and is strongly encouraged.","section":"Section 5 and Conclusions"},{"comment":"The expansion in eq. (3.8) uses the bare coupling α_{s,b}, but the renormalisation of α_s is introduced only in Section 4.1. Adding an explicit pointer from eq. (3.8) to eq. (4.1) would improve readability.","section":"Section 3, eq. (3.8)"},{"comment":"The notation H^{g,(1)}_{λ,ren} = H^{g,(1)}_λ is introduced implicitly; stating this explicitly would avoid possible confusion.","section":"Section 4.1, eq. (4.5)"},{"comment":"The text says 'a few benchmark numerical values' but the tables contain only one kinematic point per channel. Please state explicitly that further points can be made available, or include additional points in an appendix.","section":"Section 5, Tables 2 and 3"}],"recommendation":"major_revision","confidential_remarks":"The main technical risk is the completeness of the IBP reduction. The authors' admission that three relations were manually added is unusual and should be probed: the editor may wish to ask for the explicit missing relations or an independent reduction check during revision. The placeholder reference [92] suggests a coordinated submission with another group; it would be useful to confirm that the two papers are truly independent and that no circularity exists in the numerical cross-check. Overall, the paper is within the scope of a QCD/hep-ph journal and the central claim is plausible, but the reproducibility gaps should be addressed before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know the headline first: this is a genuine, difficult calculation. The authors present analytic two-loop helicity amplitudes for gg→γγ and qqbar→γγ with full top-quark mass dependence. The gg channel is new; the qqbar channel was already published by Becchetti et al. [25], but having both in one consistent analytic setup is still useful. The result rests on a large IBP reduction (26,577 scalar integrals for gg, 2,289 for qqbar) and on master integrals, including the elliptic-sector ones from their own recent work [43]. They provide the bare amplitudes in an ancillary file and benchmark finite remainders around threshold.\n\nThe paper does several things well. The UV and IR structure is handled carefully, with a clear discussion of the mixed on-shell/MS renormalisation and the decoupling relation. The internal checks are honest: UV divergence cancellation, IR pole matching, and Bose symmetry all pass. The setup is described in enough detail that an expert could reproduce the reduction. The writing is straightforward and does not oversell.\n\nThe soft spots are real but not fatal. The stress-test note points at the most sensitive assumption: Section 3 admits that the unified IBP system initially returned 91 master integrals and the authors had to add three extra relations by hand because modern software overlooked some mappings. That is exactly the kind of place where a mistake would silently change finite coefficients. Their internal checks are not sensitive to this—UV/IR poles and Bose symmetry do not probe the finite rational parts. The only external finite-value check is against ref. [92], which is a placeholder. So the central claim is currently verified only by the authors' own reduction, plus a comparison that nobody can yet inspect.\n\nThat said, I do not think the stress-test concern is disqualifying. The authors noticed the missed relations, tracked them down by mapping back to individual families, and the fact that they found three suggests they were looking carefully. The independent calculation in [92] presumably exists; the placeholder is just a citation issue. Still, the finite remainders are only available on request—that is a legitimate weakness for a paper whose whole point is usability.\n\nBottom line: this paper deserves a serious referee. Send it out, but ask the authors to (1) make the finite remainders publicly available, and (2) provide stronger evidence for IBP completeness—for instance, a numerical check of the three hand-added relations or a second, independent reduction of one of the channels. I would bring it to a reading group if someone is working on two-loop amplitudes with massive quarks, and I would cite it once the external comparison is citable and the finite remainders are actually shipped.","headline":"A real two-loop analytic amplitude, credible but with a genuine IBP-completeness caveat; deserves refereeing if the authors will share the finite remainders and address the missed-relations issue.","tokens_in":17637,"tokens_out":1858,"would_cite":true,"duration_ms":20234,"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":"Two-loop diphoton amplitudes now keep full top-quark mass dependence analytically.","keywords":["two-loop amplitudes","helicity amplitudes","diphoton production","top quark mass dependence","gluon fusion","quark-antiquark annihilation","elliptic Feynman integrals","QCD corrections"],"falsifier":"Run an independent integration-by-parts reduction of the 166 two-loop $gg\\to\\gamma\\gamma$ diagrams with a different generator of identities and see whether it reproduces the same 173 master integrals; if any further missed relation exists, the master-integral basis would shrink and the reported benchmark finite remainders at $\\theta=\\pi/6$, $s=3$ GeV would change.","tokens_in":16665,"feed_emoji":"⚛️","tokens_out":8681,"duration_ms":73408,"temperature":0.7,"pith_summary":"This paper is trying to establish that the two-loop QCD helicity amplitudes for diphoton production, both gluon-fusion and quark-antiquark annihilation, can be computed with the full top-quark mass kept exact rather than expanded away, and that the results can be written in analytic form. The interest for a reader is that these amplitudes are the hard scattering input needed for next-to-next-to-leading-order and next-to-next-to-next-to-leading-order diphoton cross sections at hadron colliders, where the heavy-quark loop first contributes at higher orders but can be numerically significant. The calculation is the first to complete this step for the gluon-fusion channel by incorporating the recently solved non-planar master integrals with elliptic sectors, and it produces finite remainders benchmarked around the top-quark threshold.","feed_headline":"First analytic two-loop diphoton amplitudes with full top mass","feed_subtitle":"Massive quark-loop contributions at NNLO and N3LO are now expressed in analytic functions, with benchmark values near threshold.","key_machinery":"The argument is carried by a three-tier construction. First, physical projectors decompose the amplitude into a basis whose tensors match the independent helicity configurations, so the eight gluon and four quark helicity amplitudes are extracted directly. Second, all Feynman integrals are organised into two planar and two non-planar integral families, with integration-by-parts reduction yielding 173 master integrals for the $gg$ channel and 65 for the $q\\bar{q}$ channel; the set includes a non-planar topology whose elliptic-sector master integrals were the last missing analytic ingredient. Third, a unified differential-equation system for the uncrossed families puts all master integrals in one consistent representation, so the final renormalised, infrared-subtracted finite remainders can be evaluated numerically at physical kinematics.","core_discovery":"The core discovery is the first analytic two-loop helicity amplitudes for $gg\\to\\gamma\\gamma$ and $q\\bar{q}\\to\\gamma\\gamma$ that keep the full dependence on the top-quark mass inside the loop. The amplitudes are decomposed directly into helicity components using physical projectors, reduced to master integrals by integration-by-parts identities, and expressed in terms of analytic functions that include the elliptic sectors of one non-planar integral family. After renormalising the heavy-quark mass on-shell and the remaining quantities in the $\\overline{\\rm MS}$ scheme, and subtracting infrared poles through standard factorisation, the authors obtain finite remainders and provide benchmark values at physical phase-space points around the top-quark threshold. The result is checked against an independent calculation, with complete numerical agreement.","pith_inferences":["A direct extension the paper only gestures at is to use the same integral families to produce the corresponding two-loop amplitudes for top-mass-dependent dijet production, since the hard functions share the same master-integral system.","The near-threshold benchmark values provide a clean test for whether mass-expanded heavy-top approximations remain reliable for diphoton production, or whether the exact threshold structure alters the finite remainders.","The availability of analytic finite remainders should make local subtraction schemes for diphoton production practical, since the singular limits of the amplitude are no longer tied to a numerical routine.","One could also scrutinise the analytic expressions for the elliptic sectors to see whether the apparent threshold complexity can be reorganised into simpler functions, which would improve numerical speed in the physical region."],"forward_implications":["The gluon-fusion channel's two-loop helicity amplitudes are now available analytically with exact top-mass dependence, replacing numerical and semi-numerical evaluations of the loop integrals as the basis for cross-section predictions.","The quark-channel two-loop amplitudes with a heavy-quark loop are also available analytically, supplying the massive-loop ingredient for NNLO diphoton production through quark-antiquark annihilation.","The mixed renormalisation scheme and infrared-factorisation recipe convert the raw amplitudes into finite remainders, with benchmark values that future subtraction schemes can use for validation.","Because the uncrossed integral families and function basis are the same for dijet production, the same framework extends to two-loop top-mass-dependent dijet amplitudes.","These finite remainders open the way to diphoton cross-section predictions at higher orders under different subtraction schemes and to quantifying heavy-quark effects at high-luminosity hadron-collider runs."],"supporting_citations":[{"why":"supplies the analytic evaluation of the non-planar two-loop master integrals with elliptic sectors that complete the amplitude basis.","marker":"[43]"},{"why":"provides the physical-projector method used to decompose the amplitude directly into helicity components.","marker":"[32, 33]"},{"why":"implements the integration-by-parts reduction used to map the scalar integrals onto a minimal set of master integrals.","marker":"[36, 37]"},{"why":"establishes the full top-mass dependence in NNLO diphoton production that this amplitude computation feeds into.","marker":"[24]"},{"why":"presents the two-loop quark-channel form factors with heavy-quark mass dependence that are related to and cross-checked against this work.","marker":"[25]"},{"why":"give the earlier numerical and semi-numerical evaluations of the gluon-fusion channel that the analytic amplitudes supersede.","marker":"[30, 31]"},{"why":"is the independent calculation of the same two-loop amplitudes used for the numerical cross-check.","marker":"[92]"}],"fun_headline_variants":["Exact top-quark mass in two-loop diphoton amplitudes","Full top mass at two loops for diphoton production","First analytic two-loop diphoton amplitudes with full mass","Massive quark loop: two-loop diphoton amplitudes computed","Two-loop diphoton helicity amplitudes with exact quark mass"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the assumption that the integration-by-parts reduction found every linear relation among the Feynman integrals, so that the reported set of 173 and 65 master integrals is complete; the authors state that standard reduction software initially missed three such relations, which they added by hand.","fun_headline_variants_meta":{"raw":{"variants":["Exact top-quark mass in two-loop diphoton amplitudes","Full top mass at two loops for diphoton production","First analytic two-loop diphoton amplitudes with full mass","Massive quark loop: two-loop diphoton amplitudes computed","Two-loop diphoton helicity amplitudes with exact quark mass"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000585,"raw_usage":{"total_tokens":2684,"prompt_tokens":813,"completion_tokens":1871,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":429,"completion_tokens_details":{"reasoning_tokens":1786}},"tokens_in":429,"tokens_out":1871,"duration_ms":14525,"temperature":1.0,"reasoning_tokens":1786,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T05:15:42.961796+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run an independent integration-by-parts reduction of the 166 two-loop $gg\\to\\gamma\\gamma$ diagrams with a different generator of identities and see whether it reproduces the same 173 master integrals; if any further missed relation exists, the master-integral basis would shrink and the reported benchmark finite remainders at $\\theta=\\pi/6$, $s=3$ GeV would change.","supporting_citations":[],"review_version":1}