{"id":"b708866d-9392-4d5b-9db4-441909f50e0e","arxiv_id":"2505.19526","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Mockenhaupt-Mitsis-Bak-Seeger restriction exponent is optimal for all d and all 0<a,b<d, via deterministic Salem set constructions.","lead":"This paper proves that the exponent in the Mockenhaupt-Mitsis-Bak-Seeger Fourier restriction theorem is optimal in every dimension and for the full parameter range. The proof is a deterministic construction of measures on sets of points well approximated by algebraic number lattices, and it also yields Salem sets.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Full-range sharpness for b>2a is not established: Remark 1.3's parameter reduction cannot produce p0 above 4d/b, while p*(a,b,d)>4d/b.","rationale":"The reader's weakest_assumption correctly identifies Remark 1.3 and the external Mitsis proposition as the delicate step for full-range coverage. The present concern sharpens this to a concrete impossibility: for b>2a, the constraints b0≤2a0, b0>b, a0>a force p*(a0,b0,d)≤4d/b, which is strictly below p*(a,b,d). Hence the construction of Theorem 1.2 cannot produce counterexamples for p0 arbitrarily close to p*(a,b,d) from below; there is a nonempty interval of p0 values for which the claimed sharpness is not addressed. This is a purely internal gap, independent of whether Mitsis Proposition 3.1 is correct. Moreover, the paper's reliance on that proposition to assert that (A)&(B) imply b≤2a is not substantiated: the one-sided Frostman upper bound does not yield the lower bound on ball masses needed for the Knapp necessary condition p≥2d/a, and Salem measures with dimension exceeding b can satisfy (A) with small a and (B) with b>2a. The construction in Sections 3-8 appears substantial and likely establishes the result for b≤2a, but the claimed full parameter range is not proven. Because the abstract and Theorem 1.2's application explicitly claim the full range, the paper as written should be rejected or substantially revised.","tokens_in":28441,"tokens_out":30505,"duration_ms":259403,"concrete_test":"Perform the feasibility computation in Remark 1.3 for d=2, a=0.6, b=1.6, p0=5.2. Here p*(0.6,1.6,2)=5.5 and 4d/b=5. Check whether there exist a0,b0 with 0.6<a0<2, 1.6<b0<2, b0≤2a0, and 5.2<p*(a0,b0,d)<5.5. Since every feasible pair satisfies p*(a0,b0,d)≤4d/b=5, no such pair exists, demonstrating that the reduction cannot cover the listed p0 even though p0<p*(a,b,d).","verdict_should_be":"REJECT","load_bearing_attack":"Remark 1.3 attempts to cover b>2a by choosing a0>a, b0>b with b0≤2a0 and p0<p*(a0,b0,d)<p*(a,b,d). However, for any feasible pair, a0≥b0/2, so p*(a0,b0,d)=4(d-a0)/b0+2 ≤ 4(d-b0/2)/b0+2 = 4d/b0 ≤ 4d/b. Meanwhile p*(a,b,d)=4(d-a)/b+2 = 4d/b + (2-4a/b) > 4d/b whenever b>2a. Thus for every p0 in the interval (4d/b, p*(a,b,d)), no choice of a0,b0 can satisfy p0<p*(a0,b0,d). Consequently Theorem 1.2 supplies counterexamples only below 4d/b for these cases, leaving the upper part of the claimed sharp range unproved. The paper's additional assertion that (A)&(B) imply b≤2a is based on the external Mitsis [24, Proposition 3.1], whose hypotheses are not reproduced; the standard Knapp necessary condition p≥2d/a is not valid from the one-sided Frostman upper bound alone, so that assertion is doubtful. The parameter gap is independent of this external proposition and is a purely internal obstruction in the reduction.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs, for each number field K of degree d and parameters τ > 1, -d < ρ < d, a deterministic Borel probability measure μ supported on the set E(K,B,τ) of points that are well approximable by inverse ideals. The measure is shown to satisfy a ball condition with exponent a < (2d - ρ₋)/(1 + τ), Fourier decay with exponent b < 2(d - ρ₊)/(1 + τ), and to violate the extension estimate for p below p(τ,ρ,q,d). By combining this with the parameter conversion in Remark 1.3, the authors claim that the exponent p*(a,b,d) in the Mockenhaupt-Mitsis-Bak-Seeger restriction theorem is best possible for every 0 < a,b < d and every dimension d, and that the construction also yields Salem sets.","tokens_in":28695,"tokens_out":24243,"duration_ms":167289,"significance":"If the full-range claim held, this would settle the sharpness of the MMBS restriction theorem in all dimensions and the full parameter range, with a deterministic construction. The paper contains substantial new technical work: a number-field generalization of Kaufman's construction, a convolution stability lemma, separation lemmas, and detailed Fourier estimates. The deterministic nature and the Salem-set corollary are notable strengths. However, the full-range claim depends on a reduction in Remark 1.3 that has a quantitative gap and on an external proposition that is not justified as stated. The portion of the argument covering b ≤ 2a appears internally consistent and is a significant contribution in its own right.","major_comments":[{"comment":"The parameter reduction does not cover the range b > 2a. For such (a,b), p*(a,b,d) = 4(d-a)/b + 2 = 4d/b + (2 - 4a/b) > 4d/b. Any admissible choice a0,b0 in the remark must satisfy b0 ≤ 2a0 and b0 > b, so p*(a0,b0,d) = 4(d-a0)/b0 + 2 ≤ 4d/b0 < 4d/b < p*(a,b,d). Therefore no choice can satisfy p0 < p*(a0,b0,d) for p0 in the interval (4d/b, p*(a,b,d)). Consequently Theorem 1.2 supplies counterexamples only for p0 < 4d/b in the case b > 2a, and the claimed full-range optimality is not established.","section":"Remark 1.3"},{"comment":"The assertion that (A) and (B) imply b ≤ 2a via Mitsis [24, Proposition 3.1] is not justified and appears false as stated. The one-sided Frostman condition (A) gives an upper bound μ(B(x,r)) ≲ r^a, not the lower bound needed for the Knapp argument. For example, Lebesgue measure restricted to [0,1]^2 satisfies (A) with a = 1/2 and (B) with b = 2, and Theorem 1.1 with these parameters gives (R) for every p ≥ p*(1/2,2,2) = 5. The cited proposition, as stated, would force p ≥ 2d/a = 8, contradicting the theorem's consequence. Unless the proposition has additional hypotheses that are not reproduced in the manuscript, the derivation of b ≤ 2a is unsupported; independently of the parameter gap above, this affects the coverage of b > 2a.","section":"Remark 1.3"}],"minor_comments":[{"comment":"Lemma 5.2 is stated without proof and is used in Lemma 5.3 and in the proof of Proposition 8.1. Since the rest of the paper is detailed, the proof should be supplied or a precise explanation given of why it follows verbatim from the proof of Lemma 5.1.","section":"Lemma 5.2"},{"comment":"The sequence (M_k) is only required to grow 'sufficiently rapidly' or 'rapidly' without an explicit list of the growth conditions. Because these conditions are used in several inductive arguments (Lemmas 4.3, 5.1, 7.5, and 8.9), stating the quantitative conditions explicitly would improve verifiability.","section":"Section 4.3"},{"comment":"The proof of Lemma 8.2 establishes a lower bound for the truncated measures μ_l and then concludes the same bound for μ. The passage to the weak limit l → ∞ should be stated explicitly, since the estimates are uniform in l but the limit step is not written out.","section":"Lemma 8.2"},{"comment":"In the proof of Lemma 6.1, the citation for the Lévy continuity theorem repeats '[2, Section 26]' for both d = 1 and d ≥ 1; the second citation appears to be intended for [4].","section":"Proof of Lemma 6.1"}],"recommendation":"major_revision","confidential_remarks":"The gap in Remark 1.3 is serious: the advertised full-range optimality result is not proven for b > 2a. If the authors cannot extend the construction to cover b > 2a, the paper should be reframed as a sharpness result for b ≤ 2a and the title/abstract adjusted accordingly. The b ≤ 2a case, together with the deterministic Salem construction, may still be a substantial contribution, but the current manuscript does not support the full-range claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Cut to the chase: this is a serious paper with a genuinely important main theorem. The algebraic number theory machinery — ideals, different ideal, the Chinese remainder argument — is used cleanly, and the deterministic construction of measures with prescribed ball growth and Fourier decay is a real advance. The Fourier and regularity estimates in Sections 6 and 7 look carefully done, and the failure argument for the extension estimate is intricate but believable. This alone is a substantial contribution to Fourier restriction for fractal measures.\n\nBut the advertised full parameter range 0<a,b<d is not actually proven as written. The reduction in Remark 1.3 has a parameter gap for b>2a. With q fixed to 2, the threshold p*(a0,b0,d) is at most 4d/b0 which is at most 4d/b, while p*(a,b,d) is strictly larger when b>2a. So the construction only supplies counterexamples below 4d/b in that range. The authors try to close the gap by citing Mitsis to assert that (A) and (R) imply p ≥ 2d/a, hence b ≤ 2a. That proposition cannot hold with only a one-sided Frostman upper bound: a measure on a set of dimension α>a can satisfy (A) with exponent a and have Fourier decay b up to 2α, which can exceed 2a. The standard Knapp necessary condition needs a matching lower bound on ball mass. As stated, the Mitsis proposition is doubtful, and the citation does not reproduce the hypotheses.\n\nThe good news is that the fix is probably simple. In Theorem 1.2, q is arbitrary, and p(τ,ρ,q,d) = q(dτ-ρ)/((q-1)(d-ρ+)) tends to infinity as q→1+. Choosing q close to 1 would cover any p0 below p*(a,b,d) even when b>2a. So the core construction is likely capable of proving the full claim; the written reduction is deficient, not the main idea.\n\nMinor soft spots: Lemma 5.2 is stated without proof and used later; the \"sufficiently rapid\" growth of Mk is standard but not made explicit. Neither changes my overall read.\n\nWho is this for? Experts in harmonic analysis and fractal restriction. It deserves a serious referee — the main theorem is important and the gap is fixable. But a referee should insist on a corrected reduction before publication.","headline":"The main construction is strong and likely correct, but the paper as written does not prove the advertised full-range sharpness for b>2a; the gap is real and fixable by using q close to 1 instead of the cited Mitsis proposition.","tokens_in":29271,"tokens_out":15631,"would_cite":false,"duration_ms":135654,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42B10","11R04","28A78"],"pacs":[],"model":"deepseek-v4-flash","headline":"The Mockenhaupt-Mitsis-Bak-Seeger restriction theorem is sharp: for every dimension d and every 0<a,b<d there is a measure meeting the hypotheses whose extension estimate fails below the claimed exponent.","keywords":["Fourier restriction","extension estimates","Mockenhaupt-Mitsis-Bak-Seeger theorem","Salem sets","Frostman condition","algebraic number theory","Diophantine approximation","sharpness"],"falsifier":"Take a concrete number field, say an imaginary quadratic field with d=2, and compute at a sequence of scales k the ratio ||\\widehat{\\Phi_{J_k,\\eta_k}\\mu}||_{L^p}/||\\Phi_{J_k,\\eta_k}||_{L^q(\\mu)} for p slightly below p(\\tau,\\rho,q,d): the claimed divergence is quantitative in M_k, so a bounded ratio at arbitrarily large scales would disprove the construction. Separately, the reduction for b>2a is falsifiable by exhibiting a measure that satisfies (A) and (R) with p<2d/a, contradicting the cited external proposition.","tokens_in":28224,"feed_emoji":"🎯","tokens_out":7587,"duration_ms":70938,"temperature":0.7,"pith_summary":"The paper proves that the exponent p*(a,b,d)=(4d-4a+2b)/b in the Mockenhaupt-Mitsis-Bak-Seeger Fourier restriction theorem cannot be improved, for every dimension d and every pair of parameters 0<a,b<d. It does so by constructing a deterministic Borel probability measure that satisfies the two hypotheses of the theorem with prescribed exponents a and b, yet fails the extension estimate for every p<p*. The construction is supported on a limsup set of points well approximable by lattices obtained from inverses of ideals in a degree-d number field, and it also yields Salem sets of dimension 2d/(1+tau). This closes a sharpness problem left open by the original theorems and previously resolved only in dimension one or in restricted parameter ranges.","feed_headline":"Fourier restriction threshold proven sharp in all dimensions","feed_subtitle":"A deterministic measure on algebraic-number lattices breaks the extension estimate below the critical exponent p*(a,b,d).","key_machinery":"The engine is the limsup set E(K,B,tau), consisting of points x in R^d that lie within distance |N(I)|^{-(tau+1)/d} of the inverse fractional ideal $I^{{-1}}$ for infinitely many ideals I of the ring of integers in a degree-d number field K. The measure is the weak limit of products of periodic bump sums F_k built from inverse prime ideals, and a convolution stability lemma controls the Fourier transform of these products scale by scale, giving the required ball regularity and Fourier decay. Failure of the extension estimate comes from constructive interference on long grids: the inverse of a prime ideal plays the role of an arithmetic progression, and the test functions are bumps supported on one such grid. A separation lemma based on ideal norms prevents the bumps from overlapping and supplies the counting estimates needed for the ball condition.","core_discovery":"The central claim is that the exponent p*(a,b,d) is best possible on the full parameter range 0<a,b<d. Theorem 1.2 constructs, for any -d<rho<d and tau>1, a probability measure supported on E(K,B,tau) that satisfies the ball condition with exponent a<2d-rho_-/(1+tau), satisfies Fourier decay with exponent b<2(d-rho_+)/(1+tau), and admits functions f_k for which the extension ratio tends to infinity whenever p<p(tau,rho,q,d). Remark 1.3 converts these parameters into arbitrary a,b with 0<a,b<d, using the cited reduction to the case b<=2a, so the constructed measure violates the extension estimate for every p below p*(a,b,d). The same construction shows that E(K,B,tau) is a Salem set of dimension 2d/(1+tau).","pith_inferences":["The full coverage of the range b>2a depends on a cited external proposition that the paper does not reprove; a reader checking the complete claim should verify that proposition's hypotheses apply exactly as stated.","The algebraic-number-field encoding suggests a portable template: other families of lattices with controlled norms and separation should yield sharpness results for other translation-invariant fractal restriction settings.","Because the counterexamples are deterministic, they may be usable where random Salem measures are not, such as in explicit constructions in additive combinatorics or in questions requiring a fixed measure with simultaneous Frostman and Fourier-decay control.","The dimension bound 2d/(1+tau) for E(K,B,tau) parallels classical Diophantine approximation exponents, so the Salem property established here can be read as an exact Fourier-analytic counterpart of those approximation results."],"forward_implications":["The Mockenhaupt-Mitsis-Bak-Seeger exponent cannot be lowered for any 0<a,b<d, so any sharper restriction estimate for measures satisfying (A) and (B) must impose additional structure beyond these two conditions.","The construction gives deterministic Salem sets of every dimension 2d/(1+tau) in every dimension d, complementing the usual random constructions.","Sharpness holds for general L^q extension estimates: for each 1<=q<infty, the extension estimate from L^q(mu) fails below p=q(dtau-rho)/((q-1)(d-rho_+)).","At the endpoint p=p*(a,b,d) the positive theorem still holds, so the result pins down the cutoff exactly rather than merely shrinking the possible range.","The same sharpness statement, with a modified construction, holds for the element-based set Eprin(K,B,tau), matching the classical Kaufman-type well-approximable sets in higher dimensions."],"supporting_citations":[{"why":"Supplies the external Proposition 3.1 used in Remark 1.3 to reduce the parameter range to b<=2a.","marker":"[24]"},{"why":"Stated the theorem with p>p* and raised the sharpness question that this paper resolves.","marker":"[25]"},{"why":"Proved the endpoint case p=p*, the statement whose exponent is shown to be best possible.","marker":"[1]"},{"why":"Proved sharpness for d=1 and b<=a, supplying the arithmetic-progression mechanism that Section 8 generalizes.","marker":"[13]"},{"why":"Constructed the deterministic measure on Eprin(K,B,tau) with maximal Fourier decay that the present construction adapts.","marker":"[9]"},{"why":"Proved full-range sharpness deterministically in d=1; its failure-of-restriction arguments underlie Proposition 8.1.","marker":"[10]"},{"why":"Provided the first deterministic Salem sets via limsup Diophantine sets, the model for E(K,B,tau).","marker":"[19]"}],"fun_headline_variants":["Sharp Fourier restriction exponent for all dimensions and parameters","Optimal exponent in Mockenhaupt-Mitsis-Bak-Seeger theorem proved","Deterministic measures yield Salem sets and sharp restriction bound","Full-range optimality for Fourier restriction in all dimensions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The full-range result relies on a previously published lemma, not proved here, saying that any measure satisfying the ball condition and the extension estimate must have p at least 2d/a; if that lemma carries hidden extra assumptions, the paper's coverage when b>2a would not follow from its construction.","fun_headline_variants_meta":{"raw":{"variants":["Sharp Fourier restriction exponent for all dimensions and parameters","Optimal exponent in Mockenhaupt-Mitsis-Bak-Seeger theorem proved","Deterministic measures yield Salem sets and sharp restriction bound","Full-range optimality for Fourier restriction in all dimensions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0011,"raw_usage":{"total_tokens":4498,"prompt_tokens":762,"completion_tokens":3736,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":378,"completion_tokens_details":{"reasoning_tokens":3667}},"tokens_in":378,"tokens_out":3736,"duration_ms":23600,"temperature":1.0,"reasoning_tokens":3667,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:13:54.818051+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a concrete number field, say an imaginary quadratic field with d=2, and compute at a sequence of scales k the ratio ||\\widehat{\\Phi_{J_k,\\eta_k}\\mu}||_{L^p}/||\\Phi_{J_k,\\eta_k}||_{L^q(\\mu)} for p slightly below p(\\tau,\\rho,q,d): the claimed divergence is quantitative in M_k, so a bounded ratio at arbitrarily large scales would disprove the construction. Separately, the reduction for b>2a is falsifiable by exhibiting a measure that satisfies (A) and (R) with p<2d/a, contradicting the cited external proposition.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the external Proposition 3.1 used in Remark 1.3 to reduce the parameter range to b<=2a."},{"cited_title":"Mockenhaupt","cited_arxiv_id":null,"evidence_quote":"Stated the theorem with p>p* and raised the sharpness question that this paper resolves."},{"cited_title":"Bak and A","cited_arxiv_id":null,"evidence_quote":"Proved the endpoint case p=p*, the statement whose exponent is shown to be best possible."},{"cited_title":"Explicit Salem sets in $\\mathbb{R}^n$","cited_arxiv_id":"1909.04581","evidence_quote":"Constructed the deterministic measure on Eprin(K,B,tau) with maximal Fourier decay that the present construction adapts."}],"review_version":1}