{"id":"b58c240c-785f-4ff0-97b5-b1b8d7aece62","arxiv_id":"2412.20577","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A two-loop multi-Reggeon exchange contribution to 2->3 amplitudes is computed for the first time, enabling extraction of the two-loop Lipatov vertex.","lead":"This conference-proceedings talk reports the first computation of the multi-Reggeon exchange pieces of two-loop 2->3 QCD amplitudes in the multi-Regge limit, the last ingredient needed to extract the two-loop Lipatov vertex from known fixed-order results. The talk reviews why this matters: the vertex controls high-energy scattering predictions and the next-order BFKL kernel.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The two-loop Lipatov-vertex extraction assumes the leading large-N_c multi-Reggeon term is exactly absorbable into the 2->3 factorized pole; the paper asserts but does not verify this, leaving the extracted vertex potentially contaminated.","rationale":"The reader's conditional verdict is appropriately cautious. The talk is mostly a review of an established program, and the 2->2 machinery (eqs. 16-18) has been verified through four loops in prior work. The genuinely new element is the 2->3 two-loop MR result (eq. 24) and the extraction of the Lipatov vertex. I considered whether the missing derivation alone is the decisive issue; it is a transparency problem but not an argument-level flaw, since the full computation is promised in [53]. The sharper issue is the absorption step: eq. (17) in 2->2 separates pole from cut by absorbing planar MR terms into the pole parameters, and this was validated by strong consistency conditions (identical corrections across channels, vanishing at four loops). For 2->3 the analogous absorption of the N_c^2 F_fact term into (19) is asserted on the basis of large-N_c universality, but (19) contains the unknown v, providing extra freedom. If the absorbed term is actually part of the cut (or should modify c_i/c_j rather than v), the three-channel agreement would not reveal the error, because F_fact is common to all channels. The proposed test - expanding (20) in W fields and comparing the N_c^2 F_fact coefficient to the universal pole expansion - would settle the absorption separately from the vertex extraction. This does not change the conditional verdict; it identifies the precise condition that the promised derivation in [53] must establish.","tokens_in":16502,"tokens_out":13104,"duration_ms":129452,"concrete_test":"Expand the 2->3 effective amplitude (20) to the relevant order in W fields for the qq channel, isolating the octet-octet component, and compare the coefficient of N_c^2 F_fact in (24) with the O(alpha_s^2) expansion of the factorized expression (19) built from the known two-loop c_q and alpha_g of [15] plus a tree-level v. If the two coefficients differ, the leading large-N_c MR term is not fully absorbed by the universal pole parameters, and the subtraction protocol must be revised; if they match, the universality assumption behind the vertex extraction is confirmed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The extraction of the two-loop Lipatov vertex rests on the decomposition (21) of the octet-octet M^{(-,-)} amplitude into single- and multi-Reggeon pieces and on the claim that, after subtracting the subleading-in-N_c terms of (24), the remainder is exactly the factorized form (19) with the same impact factors and trajectory as in 2->2. The paper justifies this by saying the leading large-N_c multi-Reggeon terms are identical for all scattering processes and can be absorbed into (19); but identity is not sufficient. Equation (19) contains the unknown two-loop vertex v, so absorbing the universal N_c^2 F_fact term into v rather than into c_i, c_j, or alpha_g is a choice. If that absorption is not the same prescription used when the 2->2 impact factors and trajectory were extracted in [14,15], the remainder after subtraction is not the bare pole contribution, and the extracted v is contaminated. Because F_fact is universal, a wrong absorption would survive the three-channel cross-check. The preprint does not show how the leading large-N_c term is handled in the subtraction, nor does it demonstrate the 2->3 analogue of the 2->2 consistency conditions (identical corrections and vanishing at higher loops).","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reviews and extends a programme for disentangling Regge poles and Regge cuts in high-energy QCD amplitudes using rapidity evolution equations formulated in terms of Reggeon fields. It reviews the 2→2 results, where the NNLL signature-odd amplitude is decomposed into a factorized single-Reggeon pole plus non-factorizable multi-Reggeon cuts, and then reports new 2→3 results: the octet-octet component of the two-loop five-parton amplitude receives multi-Reggeon contributions given in eqs. (21)-(27), and subtracting the subleading large-N_c pieces isolates the single-Reggeon pole, from which the two-loop Lipatov vertex is extracted using the factorization formula (19). The details of the calculation are delegated to the forthcoming paper [53].","tokens_in":16748,"tokens_out":5938,"duration_ms":57252,"significance":"If the reported formulas are correct, the paper would establish the first extraction of the two-loop Lipatov vertex from existing two-loop five-parton amplitude computations. This would provide a rare analytic handle on five-point two-loop amplitudes, enable multi-Regge-kinematics predictions for higher-point amplitudes, and supply a key ingredient for the next-to-leading-order BFKL kernel. The paper's strengths include a systematic framework that has already produced concrete 2→2 results, an explicit colour decomposition in eqs. (22)-(23), and a proposed cross-check of the extraction in three scattering channels. However, because the central 2→3 formulas are not derived in the text and because the absorption of the universal large-N_c term into the factorized pole is not specified, the central claim cannot currently be verified; the paper is transparent about this by referring to [53].","major_comments":[{"comment":"The central new result, the multi-Reggeon contribution to the octet-octet component of the two-loop 2→3 amplitude, is stated without derivation; the text says 'Here we briefly summarise the final results of these calculations, delegating the details to [53].' Since [53] is not available, the formulas in (24)-(27) cannot be checked from this manuscript. The extraction of the Lipatov vertex depends on subtracting these expressions from the full two-loop amplitudes, so this unverified input is load-bearing. The paper should either include enough of the derivation, or at least a detailed consistency check such as the singularity structure against known infrared or Regge constraints, or be framed strictly as an announcement contingent on [53].","section":"§5, eqs. (24)-(27)"},{"comment":"The claim that the leading large-N_c terms 'can be absorbed into the Regge-pole factorized expression (19)' is an assumption, not a demonstrated result. Because v in (19) is unknown at two loops, the universal F_fact term, including the (N_c^2+36)F_fact coefficient in the gg channel, could in principle be absorbed into v, into the impact factors c_i/c_j, or into the trajectory factor; these choices give different extracted vertices. The manuscript does not specify the absorption prescription or prove that it matches the prescription used in the 2→2 extraction of c_i and α_g. The three-channel cross-check cannot resolve this ambiguity, since F_fact is universal and would cancel in any comparison. Without this specification, the remainder after subtraction cannot be identified as the bare factorized pole contribution.","section":"§5, after eq. (23) and after eq. (27)"},{"comment":"For the 2→2 case the paper cites explicit checks of the consistency conditions that support the pole/cut separation: the footnote lists equality of the leading large-N_c corrections and their vanishing from four loops onward, verified through four loops in [14,15]. For the 2→3 case, the paper only says that factorization-violating terms 'are expected to arise only from non-planar diagrams'. No analogous check is shown that the decomposition (21) leaves exactly the 2→2 impact factors and trajectory in (19) after subtraction. This is a load-bearing gap in the justification of the extraction.","section":"§5, paragraph after eq. (19) and footnote 1"}],"minor_comments":[{"comment":"The typeset text contains numerous typographical and OCR-style artifacts, for example 'dabbed' for 'dubbed', 'multi-Reggeonexchange', 'Reggeizedgluon', and a duplicated affiliation line for the first author; these should be corrected.","section":"Throughout"},{"comment":"The colour tensors c[R1,R2] are introduced only by reference to [53]; since [53] is not available, the notation is not self-contained. A brief definition in words or in an appendix would help the reader.","section":"§5, eqs. (22)-(23)"},{"comment":"The variables z and \\bar z are described only through the phrase 'defined such that momentum conservation is admitted'; an explicit formula such as z = -p3/p4 would remove ambiguity.","section":"§5, below eq. (19)"}],"recommendation":"major_revision","confidential_remarks":"This is a conference proceedings contribution, and for such venues the expectation of standalone verifiability may be lower. If the editors treat the paper strictly as a review/announcement of results to be published in [53], the major gaps could be reduced to presentation issues and a minor revision might suffice. My recommendation of major_revision assumes a journal standard in which the central claimed result should be checkable from the text or from accessible references."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is explicit: the two-loop multi-Reggeon exchange contribution to the octet-octet component of the odd-odd 2->3 amplitude, eqs. (22)-(28), in full color. That is the missing ingredient for extracting the two-loop Lipatov vertex from the recent five-parton two-loop amplitudes, and if correct it is a substantive step toward the NLO BFKL kernel and MRK predictions at higher points. The rest of the paper is a clear review of the rapidity-evolution framework, and the review itself is done well: the pole-versus-cut distinction, the planar/non-planar logic, and the claim that only subleading-in-N_c terms break factorization all come across cleanly.\n\nThe new formulas have the right qualitative fingerprints: the (i pi)^2 signature factor, the Bloch-Wigner dilogarithm in F_fact, and the pattern that the leading large-N_c term is universal across channels while the subleading pieces are not. The three-channel (gg, qq, qg) comparison is the obvious cross-check. This is also not a fit: the MR contributions are computed directly, so the circularity concern is mild.\n\nThe soft spot, as the stress-test note says, sits in the absorption step. The paper states that the universal N_c^2 F_fact term \"can be absorbed into the Regge-pole factorized expression (19)\" because it is universal. That is necessary but not sufficient. The absorption can go into the unknown vertex v just as well as into c_i, c_j, or alpha_g, and it must match the prescription used when those impact factors and the trajectory were extracted from 2->2 scattering in [14,15]. The text does not show that the 2->3 subtraction reproduces the 2->2 consistency conditions (identical corrections across channels, vanishing at higher loops). The universal term is exactly the sort of contribution that would survive the three-channel check even if absorbed wrongly, so this is a real gap, not a manufactured one. It may well be fixed in the promised full paper [53], but the present text does not demonstrate it.\n\nThe paper is otherwise honest about its scope: it is a proceedings talk, the derivation is delegated, and the vertex itself is not given here. For a specialist in high-energy QCD or the Regge limit, it is a useful summary and a concrete new result; for a nonexpert, it is an accessible entry point to the program.\n\nIt deserves peer review: it reports first-time formulas that can be checked, and the referee's main job is to press on the absorption prescription and the consistency across channels. Send it, but treat it as a proceedings contribution and expect the follow-up to carry the full verification.","headline":"A proceedings-style review of the Edinburgh Reggeon program that includes a genuinely new two-loop multi-Reggeon result for 2->3 amplitudes, plausibly the key to the two-loop Lipatov vertex, but the extraction's absorption prescription is asserted rather than checked.","tokens_in":17293,"tokens_out":3281,"would_cite":false,"duration_ms":35242,"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":"The paper claims that the two-loop multi-Reggeon contribution to the octet-octet channel of 2→3 QCD amplitudes is given by eqs.","keywords":["Regge limit","multi-Reggeon exchange","Regge pole and Regge cut","Lipatov vertex","rapidity evolution equations","two-loop QCD amplitudes","gluon Reggeization","five-parton scattering"],"falsifier":"Compute the two-loop Lipatov vertex independently by a direct diagrammatic Reggeon computation (e.g. the approach of refs. [35–37] of the paper) or from the full two-loop 2→3 amplitudes without the Reggeon subtraction, and compare with the vertex obtained from eqs. (19)–(27); any disagreement in the colour-subleading, non-planar part would show that the multi-Reggeon subtraction missed a contribution or that the 2→2 impact factors do not transfer unchanged to 2→3.","tokens_in":16325,"feed_emoji":"⚛️","tokens_out":11611,"duration_ms":107990,"temperature":0.7,"pith_summary":"Scattering amplitudes in the high-energy limit are organised by singularities in the complex angular momentum plane: Regge poles, which in QCD give gluon Reggeization, and Regge cuts, produced by multi-Reggeon exchange. This paper reports the first computation of the two-loop multi-Reggeon exchange contribution to the octet-octet component of 2→3 amplitudes in multi-Regge kinematics, given explicitly in eqs. (24)–(27). Because this contribution breaks the single-Reggeon factorization of eq. (19), it must be subtracted before the two-loop Lipatov vertex can be read off. The paper argues that after the subtraction, the remaining single-Reggeon pole, divided by the 2→2 impact factors and gluon Regge trajectory, yields the two-loop Lipatov vertex. This matters because the vertex is currently known only at one loop and is a key input to the next-order high-energy evolution kernel and to predictions for higher-point amplitudes.","feed_headline":"Two-loop Lipatov vertex extracted from five-parton amplitudes","feed_subtitle":"Subtracting the new multi-Reggeon terms exposes the Regge pole and yields the two-loop vertex.","key_machinery":"The central object is the Reggeon effective theory built from Wilson lines in the shockwave formalism, where the projectile is a string of Wilson lines and the weak-field expansion sources individual Reggeons; the non-linear rapidity evolution equations act purely in the transverse plane. For 2→3 scattering the amplitude is written as $\\langle \\psi_j | e^{-H L_2} a_4(p_4) e^{-H L_1} | \\psi_i \\rangle$, with the mid-rapidity gluon inserted by the annihilation operator $a_4$, and the multi-Reggeon diagrams are generated by expanding this expression in Reggeon fields. The factorization formula (19), $$$M^{{(-,-)}}$_{ij\\to i'gj'}\\big|_{\\text{1-Reggeon}} = c_i(t_1)\\, $e^{{C_A \\alpha_g(t_1)L_1}}$\\, v(t_1,t_2,$p_4^{2}$)\\, $e^{{C_A \\alpha_g(t_2)L_2}}$\\, c_j(t_2)\\, $M^{{\\rm tree}}$,$$ defines the Lipatov vertex $v$ as the irreducible reggeon-reggeon-gluon emission vertex at mid rapidity. The new explicit results are the colour factors (22)–(23) and the functions (25)–(27), built from the dilogarithm combination $D_2(z,\\bar z)$ defined in eq. (28), which encode the two-loop multi-Reggeon contribution in the octet-octet channel.","core_discovery":"At two loops the signature-odd, octet-octet component of the 2→3 amplitude in multi-Regge kinematics receives, in addition to the single-Reggeon exchange, three types of multi-Reggeon contribution: $RgR^3$, $R^3gR$ and $R^3gR^3$ (eq. (21)). The paper computes the sum of these contributions in the shockwave/Reggeon effective description and finds eq. (24), with the process-dependent functions $F_{\\rm fact}$, $F_{qq}$, $F_{qg}$ given in eqs. (25)–(27); the leading large-$N_c$ terms are universal and factorizable, while the subleading terms are non-universal and break Regge-pole factorization. Subtracting these multi-Reggeon contributions from the full two-loop amplitudes isolates the single-Reggeon pole; applying the factorization formula (19) then determines the two-loop Lipatov vertex $v(t_1,t_2,p_4^2)$, using the same impact factors and gluon trajectory that appear in 2→2 scattering. The extraction can be performed in the $gg$, $qq$ and $qg$ channels, providing an internal consistency check. The paper presents this as a progress report; the complete derivation of the vertex is announced for a forthcoming companion publication.","pith_inferences":["If the $gg$, $qq$ and $qg$ extractions agree, the same subtraction protocol could be pushed to higher logarithmic accuracy, where five-Reggeon exchanges would enter and the structure of the cut contribution could be tested against soft-anomalous-dimension predictions.","The explicit dilogarithmic form of the multi-Reggeon functions suggests that the two-loop Regge cut shares transcendental structure with other transverse-momentum integrals; this could serve as a diagnostic for separating pole and cut pieces in numerical five-point computations.","A direct, independent diagrammatic Reggeon computation of the two-loop Lipatov vertex would settle the universality assumption; if it disagrees, the 2→2 impact factors would need two-loop corrections specific to 2→3 kinematics."],"forward_implications":["The two-loop Lipatov vertex follows from already-computed two-loop 2→3 amplitudes; no new full five-point computation is needed.","Performing the extraction separately for $gg$, $qq$ and $qg$ channels yields an internal check of the factorization formula (19).","With the two-loop vertex in hand, the multi-Regge limit of higher-multiplicity amplitudes can be predicted at the corresponding logarithmic accuracy.","The separation into universal large-$N_c$ factorized terms and non-universal subleading terms confirms that Regge cuts first appear in the odd-odd signature octet-octet component at two loops, as expected from the 2→2 analysis."],"supporting_citations":[{"why":"Supplies the weak-field expansion of Wilson lines in which the field $W$ sources a single Reggeon, the basis of the Reggeon effective description used throughout.","marker":"[8]"},{"why":"Establishes the Regge-pole/Regge-cut separation in 2→2 scattering and fixes the impact factors and three-loop trajectory reused by the 2→3 factorization.","marker":"[15]"},{"why":"Provides the three-loop gluon scattering amplitudes from which the gluon Regge trajectory in the factorization is determined.","marker":"[3]"},{"why":"Supplies the two-loop five-parton QCD amplitudes needed for the extraction of the vertex.","marker":"[4]"},{"why":"Provides the two-loop five-gluon amplitudes used in the $gg\\to ggg$ channel of the extraction.","marker":"[6]"},{"why":"Provides the two-loop amplitudes for the quark channels used in the $qq\\to qgq$ and $qg\\to qgg$ checks.","marker":"[7]"},{"why":"Defines the reggeon-reggeon-gluon vertex at one loop, the object whose two-loop counterpart is extracted here.","marker":"[48]"},{"why":"Announces the companion paper with the full derivation of the two-loop Lipatov vertex, of which this talk presents the multi-Reggeon results.","marker":"[53]"}],"fun_headline_variants":["Subtract Regge cuts, get two-loop Lipatov vertex","Two-loop Lipatov vertex from five-parton amplitudes","Separating Regge cuts reveals two-loop Lipatov vertex","Five-parton amplitudes yield two-loop Lipatov vertex","Regge cuts removed: two-loop Lipatov vertex exposed"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the impact factors and the gluon Regge trajectory appearing in the 2→3 factorization formula are exactly the same objects as in 2→2 scattering, and that the computed multi-Reggeon diagrams are the only two-loop contributions to the octet-octet component beyond the single-Reggeon pole.","fun_headline_variants_meta":{"raw":{"variants":["Subtract Regge cuts, get two-loop Lipatov vertex","Two-loop Lipatov vertex from five-parton amplitudes","Separating Regge cuts reveals two-loop Lipatov vertex","Five-parton amplitudes yield two-loop Lipatov vertex","Regge cuts removed: two-loop Lipatov vertex exposed"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000265,"raw_usage":{"total_tokens":1632,"prompt_tokens":994,"completion_tokens":638,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":610,"completion_tokens_details":{"reasoning_tokens":553}},"tokens_in":610,"tokens_out":638,"duration_ms":5655,"temperature":1.0,"reasoning_tokens":553,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T23:16:53.155091+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the two-loop Lipatov vertex independently by a direct diagrammatic Reggeon computation (e.g. the approach of refs. [35–37] of the paper) or from the full two-loop 2→3 amplitudes without the Reggeon subtraction, and compare with the vertex obtained from eqs. (19)–(27); any disagreement in the colour-subleading, non-planar part would show that the multi-Reggeon subtraction missed a contribution or that the 2→2 impact factors do not transfer unchanged to 2→3.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the reggeon-reggeon-gluon vertex at one loop, the object whose two-loop counterpart is extracted here."},{"cited_title":"Abreu, G","cited_arxiv_id":null,"evidence_quote":"Announces the companion paper with the full derivation of the two-loop Lipatov vertex, of which this talk presents the multi-Reggeon results."}],"review_version":1}