{"id":"39d0bd66-cb02-4f06-99fc-553af542adb2","arxiv_id":"2506.18657","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The authors give sharp uniform Fourier decay and weighted L2 restriction for all quadratic manifolds, and an almost complete implication diagram among nondegeneracy conditions.","lead":"This paper proves sharp uniform Fourier decay estimates and weighted L2 restriction bounds for quadratic manifolds of any dimension and codimension, and maps out how nine nondegeneracy conditions in harmonic analysis relate to each other. A generalist will care because the relation diagram shows which curvature assumptions are equivalent, which are strictly stronger, and which tools (restriction, decoupling, Fourier dimension) stand or fall together.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The omitted proof of (4.29) is the load-bearing gap: the narrow-case induction in Proposition 4.2 closes only if those rescaling relations hold, and the paper explicitly declines to prove them.","rationale":"The reader's weakest assumption identifies the same load-bearing gap: Proposition 4.2 is proved by induction on scales, and the narrow case closes only through the parameter relations (4.29), whose proof is omitted. This is genuinely load-bearing because the induction hypothesis for the rescaled problem involves tilde_gamma_1 and tilde_M_1, and without (4.29) the final comparison to gamma and M fails. The manuscript explicitly flags the omission, so it is not an artifact of the review pipeline. I found no stronger threat: Section 3 appears self-contained and convincing, the algebraic characterization of d_{d,1}(Q) is carefully argued, and the counterexample verifications in Sections 5 and 7, while often left to the reader, are secondary to the central claim. The right verdict remains CONDITIONAL: the paper is serious and likely correct, but this particular induction step needs to be supplied or replaced by a full derivation in arbitrary codimension.","tokens_in":82219,"tokens_out":2919,"duration_ms":35106,"concrete_test":"Take the model Q=(xi_1^2, xi_2^2) in d=n=2, choose k=3 and a dyadic K=R^delta, and independently verify (4.29) from the definitions: write down the old K^2-cubes B, the narrow sets produced by Lemma 2.2, the rescaled coordinates F on the page containing (4.26), and count the tilted boxes S. Compute tilde_gamma_1 in terms of gamma and K_1, and check whether tilde_gamma_1 times tilde_eta is at most C gamma K_1^{alpha+2} without extra assumptions. If the relation fails for some admissible tilde_eta, the induction step is invalid; if it holds, write out the omitted argument and insert it in the proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of the Du--Zhang bound (1.14) rests on the inductive Proposition 4.2. In the narrow case, after rescaling and dyadic pigeonholing, the induction closes by using the parameter relations (4.29): tilde_mu/#B is at most (log R)^6 tilde_M1 tilde_eta/M, and tilde_gamma_1 tilde_eta is at most gamma K_1^{alpha+n}. The paper states \"We omit the proof of (4.29)\" and refers to [21, Section 3] and [12, Section 3]. This is not a cosmetic omission: the second relation controls how the fractal density parameter gamma transforms under the anisotropic rescaling F. If the exponent alpha+n or the dependence on K_1 is different, or if the relation requires an extra transversality hypothesis coming from the theta-uniform condition, then the estimate (4.28) does not reduce to the required w(p,k,alpha,Q), and the induction step for Proposition 4.2, hence Theorem 1.1(1.14), is not established. The cited papers are not the same setting: [21] treats codimension 1 and [12] treats only d=3,n=2. The adaptation to arbitrary codimension n is exactly what needs proof, and the manuscript itself flags this as omitted. Since (1.14) is also used to prove sharpness of the range of the third bound in (1.13), the gap affects both the upper bound and part of the claimed optimality.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies weighted L2 restriction estimates for quadratic manifolds S_Q of arbitrary dimension d and codimension n. The first main result, Theorem 1.1, gives upper bounds for the weighted restriction exponent s(α,Q): the first and second bounds in (1.13) are derived from a new uniform Fourier decay estimate, and the bound (1.14) is obtained by adapting the Du–Zhang broad–narrow argument with lower-dimensional ℓ2Lp decoupling. The authors prove that the exponent d_{d,1}(Q)/2 in the uniform decay estimate is optimal (Theorem 3.1 and Corollary 3.8), and they show that the ranges of the first and third estimates in (1.13) are sharp. The second main result, Theorem 1.2, establishes a large diagram of implications among several nondegeneracy conditions for quadratic manifolds: Salem property, best ℓ2Lp and ℓpLp decoupling, good manifolds, the (CM) condition, best Stein–Tomas inequalities, well-curvedness, and ordinary nondegeneracy. The paper also classifies all quadratic manifolds with d+n≤5 and computes the resulting weighted restriction bounds.","tokens_in":82509,"tokens_out":3969,"duration_ms":43095,"significance":"If the results are correct, they represent a substantial advance. The uniform Fourier decay theorem in Section 3 is the most complete result of its kind for quadratic manifolds of arbitrary codimension, and the algebraic bridge m=d_{d,1}(Q) in Lemma 3.5 is clean and convincing. The weighted restriction result (1.14) extends the Du–Zhang method well beyond the codimension-one and (d,n)=(3,2) settings treated earlier, and the sharpness analysis of the first and third ranges in (1.13) is valuable. The implication diagram in Theorem 1.2 gives a useful organizing framework for many previously unrelated nondegeneracy conditions, and the paper explicitly identifies which arrows are new, which are imported from the literature, and where strictness is known. The Section 3 proof is self-contained and does not rely on fitted parameters or numerical computations. However, one load-bearing gap in the Du–Zhang induction, the omitted proof of the parameter relations (4.29), currently prevents the bound (1.14) from being fully established.","major_comments":[{"comment":"The induction step of Proposition 4.2 closes only through the two rescaling relations in (4.29), namely the bounds for tilde_mu/#B and tilde_gamma_1 tilde_eta in terms of M, gamma, K_1 and eta. The manuscript explicitly says \"We omit the proof of (4.29)\" and refers to [21, Section 3] and [12, Section 3]. This is not a cosmetic omission: the second relation controls how the fractal density parameter gamma transforms under the anisotropic rescaling, and the first relation is needed to reduce (4.28) to the claimed power R^{w(p,k,alpha,Q)}. The cited references are not in the same setting, since [21] treats codimension one and [12] treats only d=3, n=2, while the adaptation to arbitrary n is precisely what must be proved here. Until (4.29) is proved in the present generality, the estimate (1.14) and the second part of Theorem 1.1 should be regarded as conditional.","section":"§4.2, Eq. (4.29)"},{"comment":"The recovery of the parabolic weighted restriction bounds uses the lower bound X(Q,k,m) ≥ m(k−1)/k, introduced with the sentence \"repeating the arguments of Appendix C in [25]\". The paper does not state this estimate as a lemma, give its proof, or quote a precise result from [25]. Since this bound is used to derive the third estimate in (6.1), the claim that Theorem 1.1 recovers (1.4) depends on an unstated external argument. The same issue affects Corollary 6.2. Please either provide a self-contained proof of the needed X-lower bound or cite the exact statement in [25] that implies it.","section":"§6.1, proof of Corollary 6.1"}],"minor_comments":[{"comment":"In the optimality sentence, the quantity |eEQf(x)| appears even though f has not been defined in the statement; it should presumably be |eEQ1(x)|, matching the proof and Corollary 3.8.","section":"Theorem 3.1 statement"},{"comment":"In the lower-bound estimate for the radial integral the text writes \"for each θ∈S^{d−1}\"; here θ ranges over S^{n−1}, so the exponent should be n−1.","section":"§7.1, proof of implication 1"},{"comment":"The displayed range in (4.10) is strict, 0<α<d_{d,1}(Q)/2, while the theorem states a closed range at the endpoint. The endpoint follows from the second bound in (4.11), but this should be stated explicitly to avoid confusion.","section":"§4.1, Eq. (4.10)"},{"comment":"The paper is long and the proof of Theorem 1.2 is split into many subsections, so a short table or index of which subsections prove which numbered arrows (1-11) would improve readability.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The core Fourier-decay material in Section 3 is strong and likely correct, and the diagram in Theorem 1.2 is a useful contribution. The decisive issue for acceptance is the omitted proof of (4.29): this relation is load-bearing for (1.14), and the cited sources do not cover arbitrary codimension. Since the authors themselves flag the omission, I do not view this as a deliberate concealment, but the proof should be supplied before the result can be considered established. The X-lower-bound issue in Section 6.1 is also worth fixing, though it affects corollaries rather than the main theorem."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know. First, this is a serious harmonic analysis paper, and the main new results are real: the sharp uniform Fourier decay d_{d,1}(Q)/2 for arbitrary codimension (Theorem 3.1/Cor 3.8) and the almost complete relation diagram (Theorem 1.2) are both substantial and well argued. Second, the proof of the Du–Zhang bound (1.14) has a genuine gap that the authors themselves flag: equation (4.29), the rescaling relations between (M,γ) and (M1,γ1), is stated without proof and with only a pointer to [21] and [12]. That relation is what closes the narrow-case induction in Proposition 4.2. The cited papers are codimension 1 or d=3,n=2, so the adaptation to arbitrary codimension n is not automatic. This is not a cosmetic omission.\n\nNow the credit. Section 3 is the strongest part: the eigenvalue perturbation argument that converts Banner's non-uniform decay into a uniform estimate with the algebraic invariant d_{d,1}(Q) is clean, self-contained, and genuinely new. The diagram in Section 7 pulls together a lot of scattered notions, and the strictness examples and the d+n≤5 classification are useful. The paper is long, but the structure is clear and the writing is honest—they tell you when they omit something.\n\nThe soft spots, in proportion. The omitted (4.29) is the main one. It affects the proof of (1.14) and any corollaries that rely on it (e.g., Corollary 6.8). The optimality of the third range in (1.13) is argued separately via (4.31), which is less dependent, so I would not call that part blocked. Some X-estimates are imported by reference from [25], and a few counterexample verifications are left to the reader—those are minor. The reader's report says 'conditional,' and I agree: if (4.29) can be supplied, or if the paper is revised to make (1.14) explicitly conditional on an external lemma, the main results hold up.\n\nWho should read it: anyone working on weighted restriction, decoupling, or curvature conditions for submanifolds in harmonic analysis. It deserves a serious referee. I would send it to review, with an explicit request that the referee scrutinize Section 4.2 and the status of (4.29).","headline":"A substantial and mostly sound paper on weighted restriction and nondegeneracy for quadratic manifolds, with one load-bearing omission in the Du–Zhang induction (4.29) that needs to be filled or clearly conditionalized.","tokens_in":83057,"tokens_out":3223,"would_cite":true,"duration_ms":30246,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42B20","42B37"],"pacs":[],"model":"deepseek-v4-flash","headline":"For any quadratic manifold, the sharp uniform Fourier decay of its surface measure is $(1+|\\zeta|)^{-d_{d,1}(Q)/2}$, where $d_{d,1}(Q)$ is an algebraic rank invariant; the same invariant drives weighted $L^2$ restriction bounds and a…","keywords":["quadratic manifolds","weighted L2 restriction","uniform Fourier decay","decoupling inequalities","nondegeneracy conditions","broad-narrow analysis","Fourier dimension","geometric invariant theory"],"falsifier":"For a concrete quadratic manifold such as $Q=(\\xi_1^2,\\xi_1\\xi_2)$ in $d=2,n=2$, run the broad-narrow analysis at a fixed dyadic scale $R$, compute the quantities $M,\\gamma,M_1,\\gamma_1$ on both sides of (4.29), and check whether the inequality holds; failure for any $R$ would invalidate the induction behind (1.14).","tokens_in":82003,"feed_emoji":"📐","tokens_out":8000,"duration_ms":74275,"temperature":0.7,"pith_summary":"This paper establishes that the Fourier transform of the surface measure on any quadratic manifold of dimension $d$ and codimension $n$ decays uniformly at rate $(1+|\\zeta|)^{-d_{d,1}(Q)/2}$, and that this rate cannot be improved. Here $d_{d,1}(Q)$ is an algebraic invariant: the minimum rank of a linear combination of the Hessians of the quadratic forms defining the manifold. From this decay, together with the weighted layer-cake and annulus decomposition of [47] and a broad-narrow induction on scales, the paper derives weighted $L^2$ restriction estimates for every quadratic manifold, with the zero-range and saturation-range optimal. It also maps out which nondegeneracy conditions, such as Fourier decay, the Salem property, decoupling, Stein-Tomas bounds, and well-curvedness, imply which others, showing that in higher codimension the conditions genuinely diverge.","feed_headline":"All quadratic manifolds get sharp uniform Fourier decay","feed_subtitle":"An algebraic rank invariant sets the decay rate and controls weighted L2 restriction.","key_machinery":"The load-bearing object is the algebraic invariant $d_{d,1}(Q)$, defined as the minimum, over all linear changes of variables and all nonzero linear combinations of the component quadratic forms, of the number of variables that actually occur; equivalently, it is the minimum rank of the Hessian combination $Q(\\theta)$ as $\\theta$ runs over the unit sphere. It sets the uniform decay rate and equals the Fourier dimension of $S_Q$. The proof machinery also uses the decoupling exponents $\\Gamma^d_{q,p}(Q)$, the transversality quantity $X(Q,k,m)$, a $\\theta$-uniform $k$-linear restriction estimate, the weighted layer-cake/annulus decomposition of [47], and a broad-narrow induction with lower-dimensional $\\ell^2 L^p$ decoupling. For the relation diagram, the Newton-type polyhedra and the canonical affine measure from [29] carry the implications.","core_discovery":"The central discovery is a complete characterization of the sharp uniform Fourier decay for quadratic manifolds of arbitrary codimension. Writing $Q(\\theta)=\\sum_j \\theta_j \\nabla^2 Q_j$, the number of nonzero eigenvalues of $Q(\\theta)$ is locally stable in $\\theta$, and its minimum over $\\theta$ equals $d_{d,1}(Q)$. The paper converts this into the uniform bound $|\\widetilde{E_Q}1(x)| \\lesssim (1+|x|)^{-d_{d,1}(Q)/2}$, with a matching lower bound along a worst direction, so the exponent is optimal. The same quantity then controls the weighted $L^2$ restriction exponent $s(\\alpha,Q)$: the paper proves $s(\\alpha,Q)\\le 0$ for $\\alpha\\le d_{d,1}(Q)/2$, the linear bound $(2\\alpha-d_{d,1}(Q))/4$ in the middle range, and $n/2$ in the top range, with the first and third ranges sharp; a further refinement via $k$-linear restriction and lower-dimensional decoupling gives the bound (1.14). A second theorem assembles the relation diagram among nondegeneracy conditions, with strict implications such as \"best Stein-Tomas implies best $\\ell^p L^p$ decoupling.\"","pith_inferences":["Editorial extension: if the omitted relations (4.29) in the narrow-case induction admit a direct counting proof, the refined bound (1.14) would be robust for all $Q$; a counterexample to (4.29) would leave (1.14) unproved for that $Q$.","Editorial extension: the equality Fourier dimension $= d_{d,1}(Q)$ suggests a recipe for constructing sets with prescribed Fourier dimension by tuning Hessian ranks, which may be useful in geometric measure theory.","Editorial extension: the strictness pattern of the diagram suggests that as codimension grows, the correct notion of curvature depends on the operator; one testable question is whether well-curvedness is equivalent to best Stein-Tomas in the first open higher-codimension cases."],"forward_implications":["Every quadratic manifold now has an explicit weighted $L^2$ restriction bound, and the ranges where the exponent is $0$ or $n/2$ cannot be enlarged.","The Fourier dimension of $S_Q$ equals $d_{d,1}(Q)$; in particular, many higher-codimensional quadratic manifolds are not Salem even when they are smooth and curved.","Best possible Stein-Tomas restriction implies best possible $\\ell^p L^p$ decoupling for quadratic manifolds.","The (CM) condition, an integrability condition on $\\det(Q(\\theta))$, suffices for best $\\ell^p L^p$ decoupling.","If $d_{d,1}(Q)=d$, then $Q$ is nondegenerate, so a single strong algebraic condition forces many other $d_{d',n'}(Q)$ to be large."],"supporting_citations":[{"why":"Supplies the broad-narrow induction on scales and the fractal $L^2$ restriction method used to prove the refined bound (1.14).","marker":"[21]"},{"why":"Gives the formula for decoupling exponents $\\Gamma^d_{q,p}(Q)$ and defines the quantities $d_{d',n'}(Q)$ that underlie the sharp Fourier decay and the nondegeneracy conditions.","marker":"[34]"},{"why":"Provides the $\\theta$-uniform condition, the transversality quantity $X(Q,k,m)$, and the $k$-linear restriction estimates used in the broad case of the induction.","marker":"[25]"},{"why":"Contributes the weighted layer-cake and annulus decomposition that converts uniform Fourier decay into weighted $L^2$ restriction bounds.","marker":"[47]"},{"why":"Introduces the (CM) condition and the Stein-Tomas-type inequalities that appear as arrows in the nondegeneracy diagram.","marker":"[43]"},{"why":"Develops the affine measure, well-curvedness, and Newton-type polyhedra characterizations that anchor several implications in Theorem 1.2.","marker":"[29]"},{"why":"Provides prior weighted restriction results for $d=3,n=2$ and an adapted broad-narrow method that the current paper improves and generalizes.","marker":"[12]"},{"why":"Establishes the hyperbolic paraboloid weighted restriction bounds that are recovered as a special case of the new theorems.","marker":"[8]"}],"fun_headline_variants":["Sharp Fourier decay for all quadratic manifolds","Uniform decay exponent now optimal for quadratic manifolds","Nondegeneracy conditions fully mapped for quadratic manifolds","Best Stein-Tomas implies best decoupling for manifolds","Quadratic manifolds: sharp decay and complete condition diagram"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper leaves unproved the exact numerical relations (4.29) that let the induction on scales close in the narrow case; if those relations fail for some quadratic manifold, the refined bound (1.14) falls apart.","fun_headline_variants_meta":{"raw":{"variants":["Sharp Fourier decay for all quadratic manifolds","Uniform decay exponent now optimal for quadratic manifolds","Nondegeneracy conditions fully mapped for quadratic manifolds","Best Stein-Tomas implies best decoupling for manifolds","Quadratic manifolds: sharp decay and complete condition diagram"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00065,"raw_usage":{"total_tokens":2985,"prompt_tokens":953,"completion_tokens":2032,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":569,"completion_tokens_details":{"reasoning_tokens":1957}},"tokens_in":569,"tokens_out":2032,"duration_ms":14279,"temperature":1.0,"reasoning_tokens":1957,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:45:24.249378+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a concrete quadratic manifold such as $Q=(\\xi_1^2,\\xi_1\\xi_2)$ in $d=2,n=2$, run the broad-narrow analysis at a fixed dyadic scale $R$, compute the quantities $M,\\gamma,M_1,\\gamma_1$ on both sides of (4.29), and check whether the inequality holds; failure for any $R$ would invalidate the induction behind (1.14).","supporting_citations":[{"cited_title":"Du and R","cited_arxiv_id":null,"evidence_quote":"Supplies the broad-narrow induction on scales and the fractal $L^2$ restriction method used to prove the refined bound (1.14)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the formula for decoupling exponents $\\Gamma^d_{q,p}(Q)$ and defines the quantities $d_{d',n'}(Q)$ that underlie the sharp Fourier decay and the nondegeneracy conditions."},{"cited_title":"Restriction estimates for quadratic manifolds of arbitrary codimensions","cited_arxiv_id":"2308.06427","evidence_quote":"Provides the $\\theta$-uniform condition, the transversality quantity $X(Q,k,m)$, and the $k$-linear restriction estimates used in the broad case of the induction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Contributes the weighted layer-cake and annulus decomposition that converts uniform Fourier decay into weighted $L^2$ restriction bounds."},{"cited_title":"Mockenhaupt","cited_arxiv_id":null,"evidence_quote":"Introduces the (CM) condition and the Stein-Tomas-type inequalities that appear as arrows in the nondegeneracy diagram."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Develops the affine measure, well-curvedness, and Newton-type polyhedra characterizations that anchor several implications in Theorem 1.2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides prior weighted restriction results for $d=3,n=2$ and an adapted broad-narrow method that the current paper improves and generalizes."},{"cited_title":"Barron, M","cited_arxiv_id":null,"evidence_quote":"Establishes the hyperbolic paraboloid weighted restriction bounds that are recovered as a special case of the new theorems."}],"review_version":2}