{"id":"6c22c560-73cd-41c2-a4ac-ef8703a40f1f","arxiv_id":"1908.09166","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper introduces small cap decouplings, proving sharp parabolic and moment-curve exponential sum estimates and a conditional cone decoupling.","lead":"This paper develops a toolbox for decoupling oscillatory integrals into very small frequency boxes, below the usual canonical scale. It proves sharp new estimates for exponential sums on the parabola and moment curve, plus a cone result conditional on a companion theorem.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The cone result (Theorem 3.6) is conditional on the reverse square function estimate in Theorem 3.5, imported from the unavailable companion paper [16]; without that external input, the abstract's claim to have solved the cone problem is not supported by this paper alone.","rationale":"The reader's weakest_assumption correctly identifies the cone theorem as conditional on Theorem 3.5 from the companion paper [16]. I agree that this is the most load-bearing concern because it affects one of the three headline problems and depends on an unavailable external result. The manuscript itself states Theorem 3.6 as an implication, which is mathematically honest, but the abstract does not carry this caveat into its claim of having solved three problems. That mismatch supports the CONDITIONAL verdict. The reader also flags the periodicity structure in Remark 8.9 for the moment-curve proof; this is a genuine omitted verification, and the review rules require flagging it. It is secondary, however, because it concerns a technical weight-comparability step that the authors sketch in some detail, whereas the cone result depends on an entire theorem whose proof is absent from this paper. I did not find a concrete mathematical error in the parabola or moment-curve arguments, and the machine-checked or independently reproducible evidence is not present, so moderate confidence remains appropriate.","tokens_in":52220,"tokens_out":9335,"duration_ms":95976,"concrete_test":"Obtain the companion paper [16] and verify whether it proves Theorem 3.5. In addition, trace the proof of Proposition 10.2 with Theorem 3.5 replaced by the best currently available cone decoupling, Theorem 2.8: if the right-hand side becomes R^{1/8}(Σ||PθG||_4^4)^{1/4} instead of the R^ε N^{1/4}(Σ||PθG||_4^4)^{1/4} used in the text, then the final exponent in (64) gains R^{1/8+ε}, which would be too weak for Conjecture 2.7. This would confirm that the cone theorem is genuinely load-bearing on Theorem 3.5 and should be labeled conditional until [16] appears.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing concern is the status of the cone result. Theorem 3.6 does not prove Conjecture 2.7 from within this paper; it proves the implication 'Theorem 3.5 implies Conjecture 2.7'. Theorem 3.5, the reverse square function estimate for the two-dimensional cone, is stated as coming from the companion paper [16], which is listed as 'to be available soon' and whose proof is not included here. Section 10 and Proposition 10.2 use Theorem 3.5 essentially: it is the step that replaces the L4 norm of G by an L4 norm of the square function, thereby removing the canonical-scale R^{1/8} loss that would otherwise appear from applying Theorem 2.8. If Theorem 3.5 is not available or turns out to be false, the proof of small cap decoupling for the cone collapses at Proposition 10.2, and the abstract's presentation of the cone as one of three solved problems is not backed by the present manuscript. This is a dependency on an unverified external result rather than an internal inconsistency, so the correct status is CONDITIONAL rather than a definitive rejection of the method.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a general toolbox for decoupling into caps whose diameter is smaller than the canonical scale, and applies it to three problems. Theorem 3.1 establishes the sharp small cap l_p decoupling for the parabola, Theorem 3.3 gives sharp exponential sum estimates for the moment curve in R^3 in the range 0 <= beta <= 3/2, and Theorem 3.6 derives small cap decoupling for the cone from the reverse square function estimate stated as Theorem 3.5. The proofs combine refined flat decoupling, wave packet decompositions, Kakeya-type and incidence estimates, and a multilinear-to-linear reduction. The appendix by Heath-Brown applies Theorem 3.3 to a fourth derivative exponential sum estimate and derives a bound for the Riemann zeta function on the line sigma = 11/15.","tokens_in":52416,"tokens_out":6340,"duration_ms":67518,"significance":"If the main results are correct, this is a substantial contribution: Theorem 3.1 answers a conjecture that resisted the standard rescaling approach, and Theorem 3.3 gives a new range of small frequency separation mean value estimates for the moment curve. The paper is also commendably transparent: Theorem 3.6 is explicitly stated as an implication, and Remark 8.9 flags a delicate periodicity point that the authors themselves leave to the reader. There is no parameter fitting or circularity in the main argument. The principal weakness is external dependence: the cone theorem is conditional on Theorem 3.5 from the companion paper [16], which is listed as \"to be available soon\" and whose proof is not included. The moment curve proof also depends on a periodicity verification that is only sketched. These issues do not invalidate the parabola part, but they mean the paper is not yet self-contained in the form claimed by the abstract.","major_comments":[{"comment":"The cone result is not proved in this manuscript. Theorem 3.6 states only the implication \"Theorem 3.5 implies Conjecture 2.7\", and Theorem 3.5 is taken from the companion paper [16], whose proof is not included and which is listed as \"to be available soon\". Proposition 10.2 uses Theorem 3.5 essentially: it replaces the L4 norm of G by the L4 norm of the square function, thereby removing the R^{1/8} loss that would otherwise appear from Theorem 2.8. If Theorem 3.5 is unavailable, the proof of the cone theorem collapses at that point. Because the abstract presents the cone as one of the three solved problems, the manuscript overstates what is established here. The revision should either include a proof of Theorem 3.5 or restate the abstract and introduction so that the cone result is explicitly conditional on an external result.","section":"§3, Theorem 3.6; §10, Proposition 10.2"},{"comment":"The periodicity structure (S2) is an essential hypothesis for the plank incidence estimate (56), and the proof that this structure can be enforced is not given. Remark 8.9 states that the verification of A_{P1,new}/A_{P2,new} in [R^{-O(epsilon)}, R^{O(epsilon)}] is \"left to the reader\", and the discussion relies on an idealized Walsh-Fourier picture. Since Proposition 8.8, and hence Theorem 8.3 for the range 0 < beta <= 1, depends on (S2), the manuscript needs a complete proof of this weight comparability and of the claim that the structure of P(i) is preserved under the modified weights.","section":"§8, Remark 8.9; §8.3, Proposition 8.8"},{"comment":"The induction step for 1/2 < alpha < 2/3 is not fully written. In Step 9 the number of trilinear E1-rich delta-cubes inside a 1/W-cube is asserted to be bounded by (N1 M-tilde / E1)^3 without derivation, and Step 14 uses a geometric average of the bounds (35) and (39) whose intermediate exponent comparisons are not checked. The non-diagonal case is omitted throughout. Because Theorem 6.6 and Lemma 6.7 underlie the range beta > 1 of Theorem 3.3, this proof needs to be completed rather than summarized.","section":"§6.2, Theorem 6.10"}],"minor_comments":[{"comment":"In the proof that Theorem 5.1 implies Theorem 3.1, the quantities l(gamma) and l(omega) are used without definition; please define them explicitly as the lengths of the corresponding frequency intervals.","section":"§5.1"},{"comment":"The entry for the companion paper [16] is listed as \"to be available soon\"; if it has appeared by the time of revision, the citation should be updated to include the publication data.","section":"References, [16]"},{"comment":"Several displayed formulas contain typographical artifacts such as \"/greaterorsimilar\" and the author name appears as \"HONG W ANG\" on the title page; the final version should be carefully proofread.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The parabola theorem appears complete and is the strongest self-contained contribution. For the cone section, the editor should verify the status of [16] before making a final decision: if Theorem 3.5 remains unpublished, the cone claim cannot be counted as a solved problem within this manuscript. The moment curve section is promising, but the periodicity verification flagged in Remark 8.9 is load-bearing and should be supplied in full."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this one for the parabola and the moment curve. The paper proves small cap decoupling for the parabola for every alpha in [1/2,1], sharp exponential sum estimates on the twisted cubic for beta up to 3/2, and builds a toolbox—refined flat decoupling, refined canonical-scale decoupling, new Vinogradov plate/plank incidences—that looks like it will be reused. The Heath-Brown appendix is a concrete payoff: it gets a new fourth-derivative estimate and a zeta bound, and the application is explicit. That part deserves a careful referee.\n\nThe main soft spot is the cone. Theorem 3.6 is not a proof of Conjecture 2.7 from this paper; it's an implication from Theorem 3.5, the reverse square function estimate, imported from the companion paper [16], listed as 'to be available soon.' Section 10 uses Theorem 3.5 essentially at Proposition 10.2 to remove the canonical-scale R^{1/8} loss. If that estimate is unavailable or wrong, the cone argument collapses at that step. The abstract, by presenting the cone as one of three solved problems without flagging the dependency, overclaims. This is a conditionality problem, not an internal contradiction.\n\nTwo smaller wobbles. The proof of Theorem 6.6 is a two-parameter induction with details omitted in Section 6.2; I couldn't fully verify every step, though the strategy is coherent. And the periodicity assumption (S2) in Section 8 is justified only by a sketch in Remark 8.9 that leaves a nontrivial verification to the reader. Also, the abstract's moment-curve sentence doesn't mention the beta <= 3/2 restriction, which is worth fixing.\n\nNone of this shakes the parabolic or moment-curve results as far as I can see. The self-citations are to earlier decoupling theorems used as black boxes, and the conditional status of the cone result is stated plainly in Theorem 3.6 itself, just not in the abstract. The paper should go to a serious referee. The referee should have the companion paper in hand, and the authors should be asked to either include the proof of Theorem 3.5 or mark Theorem 3.6 and the abstract as conditional.","headline":"Small cap decoupling for the parabola and moment curve is the real content; the cone result is conditional on a companion paper that wasn't available, and the abstract doesn't say so.","tokens_in":52981,"tokens_out":1676,"would_cite":true,"duration_ms":17665,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42B20","11L15","11M06"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves sharp small cap decouplings for the parabola, the moment curve in R^3, and, conditional on a reverse square function estimate, the cone.","keywords":["small cap decoupling","decoupling inequalities","exponential sums","Kakeya estimates","moment curve","parabola","cone","Riemann zeta function"],"falsifier":"Find a sequence $F_R$ with spectrum in the $R^{-1}$ neighborhood of the cone for which $\\|F_R\\|_{L^4(\\mathbb R^3)} / \\|(\\sum_\\theta |P_\\theta F_R|^2)^{1/2}\\|_{L^4(\\mathbb R^3)}$ is not $O_\\epsilon(R^\\epsilon)$; that would refute the imported reverse square function estimate and remove the support for Theorem 3.6.","tokens_in":51980,"feed_emoji":"📐","tokens_out":10177,"duration_ms":93404,"temperature":0.7,"pith_summary":"The paper develops a toolbox for small cap decoupling: decoupling inequalities in which the frequency boxes have diameter smaller than the canonical scale set by curvature. With this toolbox it proves the sharp small cap $l_p$ decoupling for the parabola (Theorem 3.1), sharp exponential sum estimates with small frequency separation on the moment curve in $\\mathbb{R}^3$ (Theorem 3.3), and, conditional on an imported reverse square function estimate, the small cap decoupling for the cone (Theorem 3.6). These solve problems for which earlier methods failed, because the usual parabolic rescaling no longer works at subcanonical scales. The moment-curve estimate yields an improved fourth derivative estimate for exponential sums and, in the appendix, a new bound on the Riemann zeta-function in the critical strip.","feed_headline":"Small cap decoupling delivers sharp estimates on three surfaces","feed_subtitle":"New two-step method beats rescaling failures, and the appendix improves a zeta bound.","key_machinery":"The load-bearing device is the two-step decoupling, which replaces parabolic rescaling in the main argument. Step one is a refined flat decoupling (Proposition 4.1 and Corollary 4.2): within a flat frequency box, wave packets that cluster statistically in dual boxes gain a factor $(L^2/N)^{1/2-1/p}$ over the trivial bound. Step two is a refined canonical-scale decoupling in which the standard $R$-factor is replaced by a smaller $M$-factor measuring how many fat tubes meet each square (Theorem 5.6) or fat planks meet each cube (Theorem 7.5). The two refinements are combined with Kakeya-type incidence bounds: planar tube incidences for the parabola, new multilinear incidence bounds for plates and planks adapted to the moment curve, and plank incidences for the cone. These moment-curve plates are $(\\delta,1,1)$-plates whose normals point along tangent directions of the moment curve; their restricted direction set makes them behave like planar tubes.","core_discovery":"The central discovery is that decoupling into boxes smaller than the canonical scale obeys the same sharp exponents as ordinary decoupling, and can be proved by a two-step decoupling: refine the known canonical-scale decoupling with a statistical version of flat decoupling. For the parabola, any cap of diameter $R^{-\\alpha}$, $\\tfrac12\\le\\alpha\\le1$, supports the inequality $\\|F\\|_{L^p(\\mathbb R^2)}\\lesssim_\\epsilon R^{\\alpha(1/2-1/p)+\\epsilon}(\\sum_\\gamma\\|P_\\gamma F\\|_{L^p(\\mathbb R^2)}^p)^{1/p}$ for $2\\le p\\le 2+2/\\alpha$. On the moment curve, $\\int_{[0,1]^2\\times H}|\\sum_{k=1}^N a_k e(kx_1+k^2x_2+k^3x_3)|^{12-2\\beta}\\,dx\\lesssim_\\epsilon N^{6-2\\beta+\\epsilon}$ for any interval $H$ of length $N^{-\\beta}$, $0\\le\\beta\\le\\tfrac32$. For the cone, square-like caps of dimensions $(R^{-1/2},R^{-1},R^{-1/2})$ satisfy the analogous $l_4$ decoupling, provided the reverse square function estimate is true.","pith_inferences":["The same two-step scheme may extend to other manifolds with zero or nonzero curvature wherever a refined Kakeya or incidence estimate is available; the parabola, moment curve, and cone are illustrations rather than the boundary of the method.","The uniform-in-interval version of Theorem 3.3 needed in the appendix suggests that small cap decouplings could sharpen other exponential sum estimates, such as averages of the zeta function on short intervals, without new canonical-scale decouplings.","The cone result is only as strong as the missing reverse square function estimate; if that estimate fails, the cone conjecture remains open and the additive energy corollary for cone points would need a different proof.","Because the proof avoids parabolic rescaling in the main body, it may apply to frequency boxes that are anisotropically small in one direction only, a regime the older rescaling-based decouplings could not enter."],"forward_implications":["For the parabola, every cap diameter $R^{-\\alpha}$ now has the sharp decoupling exponent; the endpoints $\\alpha=1/2$ and $\\alpha=1$ were the only previously known cases.","The moment-curve theorem gives essentially sharp $L^{12-2\\beta}$ bounds for exponential sums on frequency-separation scale $N^{-\\beta}$, for all $0\\le\\beta\\le\\tfrac32$, with complex coefficients of modulus one.","If the reverse square function estimate is valid, the cone small cap decoupling yields $\\|F\\|_{L^4(\\mathbb R^3)}\\lesssim R^{1/4+\\epsilon}(\\sum_\\gamma \\|P_\\gamma F\\|_4^4)^{1/4}$, hence the additive energy bound $E_2(\\Lambda)\\lesssim_\\epsilon|\\Lambda|^{2+\\epsilon}$ for $\\delta$-separated subsets of the cone.","The appendix turns the moment-curve estimate into a fourth derivative estimate for exponential sums and derives $\\zeta(11/15+it)\\ll_\\epsilon (|t|+1)^{1/15+\\epsilon}$."],"supporting_citations":[{"why":"Supplies the canonical-scale $l_2$ decoupling for the parabola and cone that the two-step argument refines.","marker":"[5]"},{"why":"Supplies the reverse square function estimate on which the cone theorem is conditional.","marker":"[16]"},{"why":"Provides the standard decoupling tools, flat decoupling, interpolation, and wave packet facts used throughout.","marker":"[13]"},{"why":"Provides the refined Kakeya incidence estimates for well-spaced tubes adapted in the parabola and plate arguments.","marker":"[15]"},{"why":"Contains the refined $l_p$ decoupling estimate invoked as Theorem 5.6 for the parabola.","marker":"[17]"},{"why":"Establishes the mean value theorem for the moment curve, underpinning the canonical-scale decoupling used here.","marker":"[8]"}],"fun_headline_variants":["Small cap decoupling cracks three surfaces","Tiny caps, sharp bounds: new decoupling toolbox","Subcanonical scales: decoupling sharpened","Decoupling for tiny caps: parabola, curve, cone","New decoupling method yields sharp zeta bounds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The cone half of the paper is conditional: it assumes the reverse square function estimate for the cone in $\\mathbb{R}^3$ is true, and that estimate is imported from a companion paper whose proof was not included.","fun_headline_variants_meta":{"raw":{"variants":["Small cap decoupling cracks three surfaces","Tiny caps, sharp bounds: new decoupling toolbox","Subcanonical scales: decoupling sharpened","Decoupling for tiny caps: parabola, curve, cone","New decoupling method yields sharp zeta bounds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000973,"raw_usage":{"total_tokens":4139,"prompt_tokens":953,"completion_tokens":3186,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":569,"completion_tokens_details":{"reasoning_tokens":3112}},"tokens_in":569,"tokens_out":3186,"duration_ms":24720,"temperature":1.0,"reasoning_tokens":3112,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:19:42.494115+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a sequence $F_R$ with spectrum in the $R^{-1}$ neighborhood of the cone for which $\\|F_R\\|_{L^4(\\mathbb R^3)} / \\|(\\sum_\\theta |P_\\theta F_R|^2)^{1/2}\\|_{L^4(\\mathbb R^3)}$ is not $O_\\epsilon(R^\\epsilon)$; that would refute the imported reverse square function estimate and remove the support for Theorem 3.6.","supporting_citations":[{"cited_title":"and Demeter, C","cited_arxiv_id":null,"evidence_quote":"Supplies the canonical-scale $l_2$ decoupling for the parabola and cone that the two-step argument refines."},{"cited_title":"and Zhang, R","cited_arxiv_id":null,"evidence_quote":"Supplies the reverse square function estimate on which the cone theorem is conditional."},{"cited_title":"Fourier restriction, decoupling and applications , Cambridge University Press, 2020","cited_arxiv_id":null,"evidence_quote":"Provides the standard decoupling tools, flat decoupling, interpolation, and wave packet facts used throughout."},{"cited_title":"and Wang, H","cited_arxiv_id":null,"evidence_quote":"Provides the refined Kakeya incidence estimates for well-spaced tubes adapted in the parabola and plate arguments."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Contains the refined $l_p$ decoupling estimate invoked as Theorem 5.6 for the parabola."},{"cited_title":"and Guth, L","cited_arxiv_id":null,"evidence_quote":"Establishes the mean value theorem for the moment curve, underpinning the canonical-scale decoupling used here."}],"review_version":1}