{"id":"ee846688-78ec-4fdf-9d11-6f28eaf0a6fc","arxiv_id":"2505.23948","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Extremal bulk couplings between graviton and gluon modes are computed in F-theory AdS/CFT, yielding the graviton exchange term and a complete 1/N² correlator for the D4 theory.","lead":"This paper computes the previously unknown graviton exchange contribution to gluon scattering in a class of string theory models, fixing a term in the large-N expansion of certain 4d superconformal theories. It resolves the puzzle of divergent 'extremal' couplings by showing they are signals of operator mixing, and for one benchmark theory it completely determines the correlator at order 1/N².","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central result (1.3) rests on pole/asymptotic matching of the divergent sums (3.24); the unresolved contact-term ambiguity is a real derivation gap, though it does not threaten the pole data or the p=2 localization-fixed correlator.","rationale":"I read the paper as a top-down computation whose central object is M_R in (1.3). The paper has genuinely independent support: the bulk couplings (2.28) reproduce the stress-tensor OPE coefficient (2.30), the flat-space limit of (1.3) matches the independent result (3.31), and for p=2 in the D4 theory the localization constraints (3.51)-(3.58) fix the full 1/N^2 correlator once M_R is supplied. These checks make the result very likely correct. The load-bearing weak point is exactly the regularization of (3.24), as the reader identifies: the text asserts that the divergence is 'just a constant' and that (1.3) is verified by matching poles and large-s,t,u growth, but no explicit resummation or uniqueness theorem is provided. This is a genuine derivation gap, because an entire function of sufficiently slow growth in Mellin space is only determined up to a polynomial, and the flat-space check is degenerate under such contact-term shifts. The gap affects the decomposition of the general-p correlator into M_R and M_{F4}, but it does not affect the residues of M_R, which encode the exchanged OPE data, and it does not affect the p=2 D4 claim where localization fixes the total correlator. The reader's CONDITIONAL verdict is therefore appropriate; I would not change it, though I would ask the authors to close the gap by performing the regularization explicitly or stating a precise uniqueness argument.","tokens_in":35014,"tokens_out":11107,"duration_ms":119612,"concrete_test":"Regularize (3.24) explicitly for a fixed small p, e.g. p=3: introduce a cutoff Lambda on the k and m sums, subtract the Lambda-dependent polynomial in s, t, u fitted to low-degree terms, and take Lambda to infinity on a grid of Mellin points away from poles. If the regulated sum converges to (1.3) up to a constant, the pole/growth matching is vindicated; if a non-polynomial or channel-dependent remainder survives, (1.3) is incomplete. An analytic version is to perform the m-sum with Gosper/Zeilberger and then the k-sum as harmonic sums, comparing directly to the H_{...} terms in (1.3).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The key claim that the graviton exchange Mellin amplitude is the closed form (1.3) is supported in Section 3.2 only by the assertion that the divergent double sums in (3.24) are determined, up to a constant absorbable into M_{F4}, by their poles in s, t, u and their large-s,t,u growth. This is not established: (3.24) is never explicitly regularized or resummed, and no uniqueness argument bounding the polynomial remainder is supplied. The statement 'We can regularize this divergence by checking that our final closed form answer (1.3) has the same poles and the same growth' is a consistency check, not a regularization. The flat-space match to (3.31) cannot resolve the issue, because a contact-term shift changes both sides only by a polynomial in s, t, u. Consequently, for general p the separation of M_R from the unknown M_{F4} coefficients in (1.2) is convention-dependent; only the residues of M_R, i.e. the exchanged graviton OPE data, are fixed by the summands. The D4 specialization is safer: the localization constraints (3.51)-(3.58) fix the total 1/N^2 correlator, so the p=2 claim survives even if M_R has a contact-term ambiguity. The gap is a missing proof of uniqueness or an explicit resummation of (3.24), not a demonstrated contradiction; the result is very likely correct, but the advertised determination of M_R for general p is conditional on closing this gap.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies extremal and super-extremal cubic couplings between gluon KK modes and graviton KK modes in F-theory AdS5 backgrounds, using the DBI/WZ action on the seven-branes. It then uses these couplings to compute the graviton-exchange contribution M_R to the holographic gluon correlator <22pp>, obtaining the closed Mellin-space formula (1.3). The authors use M_R to unmix single-trace graviton modes from double-trace gluon operators, thereby giving a finite interpretation of the divergent extremal Witten diagrams. For the D4 theory with p=2, they combine the new M_R with supersymmetric localization constraints to fix the full correlator at order 1/N^2 and extract CFT data and a flat-space prediction for A_F4.","tokens_in":35267,"tokens_out":5771,"duration_ms":59786,"significance":"If the central formula (1.3) is established, the paper makes a substantial contribution: it provides the first concrete top-down example with non-vanishing extremal couplings, shows that an infinite tower of graviton exchanges contributes at the same order as gluon loops, and explains the associated single-trace/double-trace mixing. The bulk-coupling computation in Appendix A is detailed and passes a nontrivial consistency check: the stress-tensor OPE coefficient (2.30) matches the known result. The D4 specialization is especially valuable because the localization constraints (3.51)-(3.58) fix the total 1/N^2 correlator, so the p=2 consequences survive even if the general-p separation of M_R from M_F4 is convention-dependent. The flat-space comparison with [24] and the matching of the off-diagonal mixing matrix with [12] are additional checks. The main weakness is that the passage from the divergent sums (3.24) to the closed form (1.3) is not derived, only checked by pole and growth matching.","major_comments":[{"comment":"The central claim that the graviton exchange Mellin amplitude equals the closed form (1.3) is not established. The double sums in (3.24) are never explicitly regularized or resummed. The statement that 'the divergence is just a constant that can be absorbed into the definition of the contact term M_F4' is asserted, but no uniqueness argument is given for the polynomial remainder. The subsequent sentence, 'We can regularize this divergence by checking that our final closed form answer (1.3) has the same poles... and the same growth,' is a consistency check, not a derivation. Without an explicit resummation or a proof that the sums are determined up to a constant by their pole data and large-s,t,u growth, the advertised determination of M_R for general p is conditional.","section":"Section 3.2, Eq. (3.24) to Eq. (1.3)"},{"comment":"The flat-space match to A_R from [24] cannot resolve the contact-term ambiguity. Under the flat-space limit formula (3.26), a polynomial ambiguity in s,t,u maps to a polynomial in the flat-space Mandelstam variables, which is precisely the ambiguity that can be absorbed into A_F4 in (3.25). Therefore the match only checks the pole residues and the logarithmic growth, not the overall normalization of the contact-term-free part of M_R. For general p, only the residues of M_R, i.e. the exchanged graviton OPE data, are fixed by the summands in (3.24); the closed form (1.3) represents a particular subtraction prescription.","section":"Section 3.2, Eqs. (3.25)-(3.31)"},{"comment":"The D4 p=2 result is robust despite the general-p gap because the localization constraints (3.51)-(3.58) fix the total correlator at order 1/N^2, including the contact-term coefficients b^i_F4. However, the paper should explicitly state that this robustness does not extend to general p: for p>2, the separation of M_R from the unknown M_F4 coefficients in (1.2) is convention-dependent unless the regularization of (3.24) is supplied. I would like the authors either to provide the missing derivation or to reframe the general-p claim as fixing only the residues and the p=2 combination.","section":"Section 3.4, Eqs. (3.51)-(3.58)"}],"minor_comments":[{"comment":"The abstract contains a grammatical error: 'such coupling are non-zero' should read 'such couplings are non-zero'.","section":"Abstract"},{"comment":"In (3.60), the coefficient of t7 is written as a^1_F4, but the definitions in (3.61) strongly suggest this should be a^3_F4.","section":"Section 3.4, Eq. (3.60)"},{"comment":"The phrase 'the holographically renormalised result (3.25) of [12]' is confusing, because (3.25) in this paper is a different flat-space amplitude. The reference should cite the equation number in [12] explicitly.","section":"Section 3.3, footnote 22"},{"comment":"For the 1-loop gluon term, the regularized expression (3.15) is presented only for p=2, and the relation of (3.15) to the divergent sums (3.14) is not explained. A brief statement about the subtraction scheme would improve reproducibility.","section":"Section 3.2, Eq. (3.14) and Eq. (3.15)"},{"comment":"In (3.24), the 'crossed expressions from the <2p2p> configuration' are not written out explicitly. Since these are needed to reproduce the t- and u-channel sums, writing them out would make the computation more transparent.","section":"Section 3.2, Eq. (3.24)"}],"recommendation":"major_revision","confidential_remarks":"The paper is strong and likely correct, but the headline result (1.3) for general p rests on an unproven regularization step. I would not reject: the D4 p=2 result and the OPE data from the residues appear solid, and the gap may be closable. However, if the authors cannot supply a derivation or a precise uniqueness statement, they should explicitly downgrade the general-p claim to a convention-dependent one."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this paper gives the first complete 1/N^2 holographic correlator for a 4d N=2 SCFT beyond maximal supersymmetry, for the D4 theory. The new machinery is the bulk cubic couplings beta_{pq k_r} in (2.28), derived from DBI/WZ, and the closed Mellin-space graviton exchange (1.3) for general p. Both look right to me. The couplings pass the stress-tensor OPE check (2.30), and the flat space limit of (1.3) matches [24] precisely. The tau-independent part of A_{F4} for D4 is a new prediction, fixed via localization; that part is on solid ground.\n\nWhat I'd flag: the passage from the divergent double sums (3.24) to the closed form (1.3) is not a derivation. The paper says that checking poles and large-s,t,u growth fixes the answer up to a constant, but no uniqueness argument is given and the sums are never explicitly regularized or resummed. The flat-space match to (3.31) can't resolve the ambiguity because a contact-term shift moves both sides by a polynomial. So for general p, the split between M_R and M_{F4} is convention-dependent; only the residues, i.e. the exchanged OPE data, are fixed by the summands. The D4 p=2 case is much safer, because the localization constraints fix the total correlator regardless of that constant. I read this as a derivation gap, not a demonstrated error. The final answers pass every external check, and the mixing matrix matches [12] independently, so I'd bet on (1.3) being correct.\n\nAlso worth noting: this is a genuinely top-down computation. No constants are fitted to the CFT amplitude; the only external input is the independent flat-space A_R from [24] (which shares authors but is a separate computation). The unmixing section is careful, and the twist-6 data appears new.\n\nWho's it for: people working on AdS/CFT correlators, F-theory holography, and the bootstrap. It deserves a serious referee. The referee should push on a proper regularization of (3.24) or at least a clear statement that the contact-term split is a convention, and the advertised ``completely fix'' should be softened accordingly. I'd engage with it.","headline":"A strong top-down computation of extremal bulk couplings and graviton exchange in F-theory AdS/CFT, with a real but narrow derivation gap in the resummation that likely does not change the answer.","tokens_in":35903,"tokens_out":1860,"would_cite":true,"duration_ms":18198,"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 a closed Mellin-space formula for graviton exchange in gluon scattering on F-theory AdS5 backgrounds, and uses it with supersymmetric localization to fix the D4 theory's ⟨2222⟩ correlator completely at order 1/N².","keywords":["AdS/CFT","extremal couplings","graviton exchange","gluon KK modes","F-theory","supersymmetric localization","Mellin amplitudes","operator mixing"],"falsifier":"Take the regularized double sum (3.24) for a specific value such as $p=4$, evaluate it with an explicit regulator on the $m$ and $k$ sums at generic complex $s,t,u$, and check that it equals the closed form (1.3) up to a constant independent of $s,t,u$; any $s,t,u$-dependent remainder would show that the graviton-exchange term is not uniquely determined for general $p$. A complementary test is an independent computation of the $D_4$ $\\langle 2222 \\rangle$ correlator at order $1/N^2$ that checks the $\\tau$-independent part of $A_{F^4}$ predicted in (3.61).","tokens_in":34751,"feed_emoji":"📐","tokens_out":14392,"duration_ms":115237,"temperature":0.7,"pith_summary":"Extremal cubic couplings—those whose dimensions satisfy Δ_i = Δ_j + Δ_k + 2a—make naive three-point Witten diagrams diverge and were previously known to vanish in maximally supersymmetric examples. This paper shows that in the F-theory AdS5 backgrounds dual to 4d N=2 SCFTs, an infinite set of such (super-)extremal couplings between gluon modes on sevenbranes and graviton KK modes are nonzero, and that their exchange diagrams sum to a finite closed expression. The graviton-exchange term so obtained completes the gluon correlator ⟨22pp⟩ at order 1/N², and is verified by matching its flat-space limit to the 8d tree-level graviton amplitude. For the D4 gauge theory the new term, together with supersymmetric localization, fixes the ⟨2222⟩ correlator completely at order 1/N², including the τ-independent part that previous work could not determine. A sympathetic reader would care because this removes a missing loop-order contribution from every correlator in this class and shows how divergent extremal couplings are resolved by operator mixing.","feed_headline":"Infinite graviton tower collapses to one gluon-scattering formula","feed_subtitle":"The new term completes F-theory gluon correlators at 1/N² and fixes the D4 case entirely.","key_machinery":"The load-bearing object is the bulk cubic coupling $\\beta_{pqk_r}$ between two gluon KK modes and a graviton KK mode, computed by expanding the DBI and Wess–Zumino actions of the sevenbranes, canonically normalizing all fields, and integrating over the internal $S^5/S^3$ geometry. Its explicit form (2.28) is a ratio of Gamma functions with the selection rule $r=|p-q|+2, |p-q|+4, \\ldots, p+q-2$, and it vanishes in the half-BPS endpoint of the extremal range; the nonzero couplings are precisely the (super-)extremal ones that make naive three-point Witten diagrams diverge. Substituting this coupling into the three-point dictionary and into the scalar-exchange block sums gives the divergent double-sum representation (3.24); the closed Mellin amplitude (1.3) is the function with the same poles and large-$s,t,u$ growth. The second mechanism is the mixing matrix $M_n$ of Section 3.3: the extremal couplings force the double-trace gluon operators $:\\!\\phi^A_p \\square^{n-p}\\phi^A_p\\!:$ to mix with the single-trace graviton operator $\\rho_{2n}$, and the four-point computation supplies the off-diagonal entries of $M_n$, converting divergent three-point functions into finite leading anomalous dimensions of the unmixed operators.","core_discovery":"On the paper's own terms, the central discovery is that graviton exchange in gluon scattering on AdS$_5 \\times S^3$, naively an infinite sum of exchange diagrams over the whole tower of KK graviton modes, sums to a single closed Mellin-space amplitude. For the correlator $\\langle 22pp \\rangle$ of the gluon superprimaries of dimensions $2$ and $p$, this amplitude reads $$M_R = -\\frac{p}{(p-2)!\\$\\Delta$}\\left[\\left((p+1)H_1 - \\frac{s}{2} + \\frac{4}{s-2}\\right)\\$delta^{{AB}}$\\$delta^{{CD}}$ + \\left((p+1)H_2 + \\frac{p-u}{2} + \\frac{2p}{u-p}\\right)\\$delta^{{AC}}$\\$delta^{{BD}}$ + \\left((p+1)H_2 + \\frac{p-t}{2} + \\frac{2p}{t-p}\\right)\\$delta^{{AD}}$\\$delta^{{BC}}$\\right],$$ with $H_x$ the relevant harmonic series. Each channel contains the simple pole expected from the exchanged tower, at $s=2$ in the direct channel and at $u=p$ or $t=p$ in the crossed channels. The coefficients are fixed by the nonzero (super-)extremal bulk couplings $\\beta_{pqk_r}$ computed from the sevenbrane action, and the closed form is checked by matching the poles and large-$s,t,u$ growth of the regularized double sum and by matching its flat-space limit to the independent 8d graviton-exchange amplitude. In the $D_4$ theory the new term supplies the missing ingredient that lets supersymmetric localization fix the contact term $M_{F^4}$ completely, including its $\\tau$-independent part, so the $\\langle 2222 \\rangle$ correlator is fully determined at order $1/N^2$.","pith_inferences":["Extension: the same bulk-coupling computation should supply graviton exchange for the more general correlators $\\langle ppqq \\rangle$ once the one-loop gluon terms are computed for arbitrary $p,q$, a step the paper leaves open.","Extension: in gluon scattering on AdS$_{d+1}\\times S^3$ for $d=3,5,6$, the graviton exchange term is the leading $1/N^2$ correction (or competes with two-loop gluon exchange for $d=3$), so the machinery developed here gives a template for unambiguous CFT data in those cases.","Extension: the explicit mixing matrices for twist four and six suggest that for generic twist $2n$ the leading spectrum comes from diagonalising an $n\\times n$ matrix $M_n$ assembled from gluon and graviton exchange data; a closed-form-in-$n$ result may be derivable.","Extension: because the flat-space graviton amplitude is sensitive to the compact $S^3$ directions, precision numerical bootstrap bounds on these 4d $\\mathcal{N}=2$ SCFTs could independently check the predicted $1/N^2$ anomalous dimensions."],"forward_implications":["The graviton-exchange term $M_R$ is now known in closed form for every $p$ in $\\langle 22pp \\rangle$, so the full correlator at order $1/N^2$ is determined once the gluon one-loop and contact terms are fixed for that $p$.","The flat-space limit of $M_R$ reproduces the 8d tree-level graviton exchange amplitude, confirming that the exchanged graviton propagates in the full 10d spacetime and that the infinite KK tower is needed for the correct logarithmic $s,t,u$ behaviour.","For the $D_4$ theory with $p=2$, the $\\langle 2222 \\rangle$ correlator is completely fixed at order $1/N^2$, and all unambiguous twist-four anomalous dimensions in (3.62)-(3.64) become predictions that can be compared with other methods.","The unmixing of graviton modes with gluon double traces gives finite leading anomalous dimensions and OPE coefficients in (3.42)-(3.43) for twist four and six, resolving the divergence of the extremal three-point diagrams.","The paper's data at twist $2n$ in principle suffice to unmix the leading CFT data for generic $n$, because the required gluon-exchange data for $\\langle ppqq \\rangle$ were already known."],"supporting_citations":[{"why":"Supplies the S5 scalar harmonics and normalization technology for the graviton KK modes used in the cubic-coupling computation.","marker":"[4]"},{"why":"Provides the IIB supergravity mass spectrum on S5 and the diagonalization of scalar fluctuations behind the graviton mode action (2.20).","marker":"[57]"},{"why":"Identifies the holographic duals of D3-branes at F-theory singularities and the gluon vector harmonics on S3 used for the gluon modes.","marker":"[16]"},{"why":"Establishes the regularized interpretation of extremal couplings as single-trace/double-trace mixing, which the paper's unmixing calculation independently matches.","marker":"[12]"},{"why":"Fixed the tree-level gluon exchange term M_F2, the leading 1/N contribution to the correlator that the new graviton term supplements at 1/N².","marker":"[20]"},{"why":"Computed the one-loop gluon exchange term M_F2|F2 at the same order as graviton exchange and supplies the B-function sums used in the analysis.","marker":"[21]"},{"why":"Provides the flat-space 8d graviton exchange amplitude and logarithmic threshold that the flat-space limit of (1.3) must match, along with the flavor projectors.","marker":"[24]"},{"why":"Used localization to fix the τ-dependent part of M_F4 for p=2; the new graviton term extends this to the full M_F4.","marker":"[23]"},{"why":"Supplies the integrated-constraint relations between derivatives of the mass-deformed sphere free energy and Mellin integrals used to fix the D4 contact terms.","marker":"[26]"}],"fun_headline_variants":["Infinite graviton towers sum to one closed Mellin amplitude","One Mellin amplitude collapses infinite graviton exchange","Graviton tower sum gives complete gluon correlators at 1/N^2","Graviton exchange sum fixes gluon scattering at order 1/N^2","Closed Mellin amplitude from graviton towers for gluon correlators"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the divergence of the double sums in (3.24) is only an $s,t,u$-independent constant that can be absorbed into the contact term $M_{F^4}$; if the regularization produced a polynomial remainder in $s,t,u$, the closed form (1.3) would not be determined uniquely, though the $D_4$ localization constraints would still fix the $p=2$ combination.","fun_headline_variants_meta":{"raw":{"variants":["Infinite graviton towers sum to one closed Mellin amplitude","One Mellin amplitude collapses infinite graviton exchange","Graviton tower sum gives complete gluon correlators at 1/N^2","Graviton exchange sum fixes gluon scattering at order 1/N^2","Closed Mellin amplitude from graviton towers for gluon correlators"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00056,"raw_usage":{"total_tokens":2813,"prompt_tokens":1251,"completion_tokens":1562,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":867,"completion_tokens_details":{"reasoning_tokens":1469}},"tokens_in":867,"tokens_out":1562,"duration_ms":10456,"temperature":1.0,"reasoning_tokens":1469,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:38:27.057206+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the regularized double sum (3.24) for a specific value such as $p=4$, evaluate it with an explicit regulator on the $m$ and $k$ sums at generic complex $s,t,u$, and check that it equals the closed form (1.3) up to a constant independent of $s,t,u$; any $s,t,u$-dependent remainder would show that the graviton-exchange term is not uniquely determined for general $p$. A complementary test is an independent computation of the $D_4$ $\\langle 2222 \\rangle$ correlator at order $1/N^2$ that checks the $\\tau$-independent part of $A_{F^4}$ predicted in (3.61).","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the IIB supergravity mass spectrum on S5 and the diagonalization of scalar fluctuations behind the graviton mode action (2.20)."},{"cited_title":"Behan, S","cited_arxiv_id":null,"evidence_quote":"Provides the flat-space 8d graviton exchange amplitude and logarithmic threshold that the flat-space limit of (1.3) must match, along with the flavor projectors."},{"cited_title":"Behan, S","cited_arxiv_id":null,"evidence_quote":"Used localization to fix the τ-dependent part of M_F4 for p=2; the new graviton term extends this to the full M_F4."}],"review_version":1}