{"id":"5934030d-523c-455f-b80b-673fdca0a531","arxiv_id":"2602.17613","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The weighted L^p(|x|^alpha) type set of the spherical maximal operator with dilation set E is characterized, up to endpoints, by the Legendre-Assouad function of E.","lead":"For spherical averages over a prescribed set of radii, the exact power weights |x|^alpha for which the maximal operator is bounded on weighted L^p are now characterized by a fractal-dimension function of the radius set. The result closes, apart from endpoint cases, a weighted-inequality problem open since 2002.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified","rationale":"I stress-tested the central claim by following the two proof directions. In the lower-bound direction, I verified that Lemma 4.1's two cases correctly yield (4.2), and that the supremum over R, j, I matches the definition of ν# after taking logarithms; the algebra for the |I| exponent and the 2^{-j} exponent is internally consistent. In the upper-bound direction, I checked that Proposition 5.1's hypothesis (5.2) is exactly what the global condition (5.1) produces after scaling, and that the sum over k in the final display cancels the k-dependence from N^{1/p} against the hypothesis, leaving only (1+j)2^{-jε/p}A. I also rechecked the interpolation logic behind Lemma 5.3: the correct route is interpolation between p=1 with weight |x|^α and p=∞ unweighted, which yields the claimed factor 2^{-j(d-1+α)/p}; the earlier apparent mismatch with the α=1-d and α=0 estimates disappears when the p=∞ endpoint is included. Similarly, the derivative estimates in (5.15b) are consistent with the shell-measure bound at p=1 and the Fourier decay (5.11b) at p=2. The only genuine soft spot is that several of these estimates are presented with sketches rather than full proofs, but I could not identify a concrete error or a missing assumption. The reader's weakest-assumption identification points to the same region, but I do not find it sufficient to move the verdict; hence UNCHANGED.","tokens_in":14483,"tokens_out":61893,"duration_ms":489580,"concrete_test":"Re-derive (5.6) and (5.15b) explicitly. For (5.6), verify the p=1 endpoint via sup |K^j_t|=O(2^j) and ∫_{B_j}|x|^α dx≈2^{-j(d+α)}, then interpolate with the unweighted p=∞ endpoint to confirm the exponent (d-1+α)/p. For (5.15b), compute the p=2 derivative operator norm from (5.11b) and check the p=1 shell-measure argument gives exactly the factor 2^{-k(d-1)}. If either exponent differs, Proposition 5.1—and hence the sufficiency half of Theorem 1.1—would require revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I could not find a concrete load-bearing error in Theorem 1.1. The necessary-condition extraction from (4.2) to (4.1), the scaling argument through E_R, the two-case Knapp lower bounds in Lemma 4.1, and the k-dependence cancellation against hypothesis (5.2) all check out. The sufficiency direction is least formalized at the frequency-localized kernel estimates (5.9a,b)-(5.11a,b) and the derivative estimate (5.15b), which are asserted with sketches ('analogous'). If any of these carried a wrong decay rate or a wrong interpolation endpoint, Proposition 5.1 would fail. But the stated rates are consistent with standard oscillatory-integral heuristics: (5.11a) uses the moment condition N0≥(d-1)/2, (5.11b) loses one power of 2^j as expected, and the p=1 endpoint of (5.15b) follows from the shell measure O(2^{-k(d-1)}2^{-j}) times sup |dK/dt|=O(2^{2j}). I found no internal inconsistency in the algebra or the exponent accounting.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the spherical maximal operator M_E with a general set of dilation radii E, acting on weighted L^p(|x|^alpha) spaces. The main result, Theorem 1.1, gives a complete description — up to endpoint cases — of the closure of the type set T_E = {(1/p, alpha/p) : M_E bounded on L^p(|x|^alpha)}. This description is in terms of the Legendre–Assouad function of E, with the region bounded by the two curves U(p) and L(p). The proof has two parts: in Section 4, two families of Knapp-type test functions are used to extract the necessary entropy condition, reducing it to the lower bound L(p) <= alpha; in Section 5, a frequency-localized decomposition of the spherical averaging kernel is used to prove sufficiency under the scale-invariant hypothesis (5.2). The paper also discusses Assouad regular sets and finite unions of such sets.","tokens_in":14472,"tokens_out":30931,"duration_ms":230844,"significance":"If correct, this settles, up to endpoint issues, an open problem of Duoandikoetxea and Seijo and gives the first complete weighted description for arbitrary dilation sets E. The result unifies the unweighted theorem of Seeger–Wainger–Wright with the earlier partial weighted results of Duoandikoetxea–Vega and Duoandikoetxea–Seijo. The lower-bound construction is particularly elegant: the two cases of Lemma 4.1 yield the exact entropy quantity appearing in the Legendre–Assouad function, and the exponent bookkeeping is internally consistent. The sufficiency proof, while relying on several standard frequency estimates, is organized so that the role of the entropy hypothesis (5.2) is transparent. The paper is a substantial contribution to the harmonic analysis of maximal operators with restricted dilations.","major_comments":[{"comment":"The estimate (5.15b) is load-bearing: it controls the derivative term in Lemma 5.4 and hence is needed for Proposition 5.1. The proof is omitted with the words 'completely analogous'. Since the statement is not a routine off-the-shelf estimate (it involves interpolation between p=1 and p=2 and an endpoint where the L^1 norm of dK^j_t/dt contributes), the authors should include the short argument or give a precise reference. I checked that the stated rate is consistent: for p=2 it follows from (5.11b) together with the 2^{-j} factor, and for p=1 from the shell-measure bound plus ||dK^j_t/dt||_1 = O(2^j), but the written proof should not leave this to the reader.","section":"Section 5.1, Eq. (5.15b)"},{"comment":"The reduction to the case alpha<0 and p <= 1+1/(d-1) is justified by saying that alpha >= 0 was already handled in [6]. The introduction, however, cites [6] as proving a positive result only for p > p_1. Since the region for alpha >= 0 and p in (p_beta, p_1] is nontrivial (the upper bound U(p) is positive there), the authors should state precisely which theorem in [6] covers this case, or provide the argument. If [6] does not cover it, the sufficiency proof would need to be extended.","section":"Section 5, first paragraph"}],"minor_comments":[{"comment":"The statement says 'Let alpha >= 1-d', but the proof interpolates between alpha = 1-d and alpha = 0 and therefore establishes the estimate only for 1-d <= alpha <= 0. Since Proposition 5.1 assumes alpha < 0, this does not affect the argument, but the statement should be corrected (or the proof extended).","section":"Lemma 5.3"},{"comment":"The displayed estimate uses 'sup_{s in E_R ∩ I}', but E_R has not been defined in the proof of Lemma 4.1; it should be 'sup_{s in E ∩ I}'.","section":"Section 4.2, after Eq. (4.7)"},{"comment":"The formula for L(p) for finite unions appears to be missing parentheses. As typeset, 'β_j γ_j −(d−1)(p−1)' is ambiguous and likely should read 'β_j(γ_j −(d−1)(p−1)) / (γ_j−β_j)', matching the formula from the Assouad regular case in Section 3.1.","section":"Section 3.2"},{"comment":"The sentence 'We may also assume that alpha < 0 since the case alpha >= 0 was already handled in [6]' would benefit from a precise theorem number or page reference in the bibliography, especially because the introduction's summary of [6] is stated for p > p_1.","section":"Section 5, first paragraph"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is a strong contribution and the central theorem appears sound; the lower-bound construction is convincing and the upper-bound machinery is standard. The main issues are local exposition points, not technical errors. I recommend minor revision, with the expectation that the authors add the missing details in (5.15b) and clarify the reference to [6] for the alpha >= 0 range."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is the real thing. It closes the weighted spherical maximal problem for arbitrary dilation sets, up to the standard endpoint caveat, and pins down the Legendre–Assouad function as the object that governs the type set. The sharp lower boundary for non-Assouad-regular sets (e.g., {1+n^{-a}}, a≠1) is genuinely new; prior work only had necessary conditions in that range and a positive result above p_1.\n\nWhat the paper does well: the necessary and sufficient conditions match exactly, the two-family Knapp constructions in §4 are clean and the exponent accounting checks out, and the reduction from global to local entropy via scaling (Lemma 2.1) is a neat way to get the global statement from compact sets. I verified the equivalence of the explicit and implicit type-set descriptions, the boundary values, and the cancellation of the k-dependence against hypothesis (5.2). Nothing in the proof looks fitted or circular.\n\nThe soft spots are real but not disqualifying. The sufficiency proof leans on frequency-localized kernel estimates (5.9a,b)–(5.11a,b) and the derivative bound (5.15b), and several of these are asserted with “analogous” or “follows by interpolation” rather than written out. The p=1 case of (5.9a) is sketched in a sentence. A referee should ask for full details there; if any decay rate were off by a power of 2^j, the converse half would collapse. But the stated rates are consistent with standard stationary-phase heuristics and the moment condition N_0 ≥ (d-1)/2 is the right one. The endpoint cases are excluded by design, and the result characterizes the closure of the type set — both honestly stated.\n\nI could not find a load-bearing error. The paper is written for harmonic analysts working on maximal operators and fractal dimensions of dilation sets; for them it is a must-read. Send it to a serious referee, with instructions to scrutinize the sketched kernel estimates.","headline":"This paper genuinely resolves the weighted spherical maximal problem for arbitrary dilation sets up to endpoints, with the Legendre–Assouad function as the governing object; the main theorem holds together, but several key kernel estimates are sketched rather than fully proved.","tokens_in":15247,"tokens_out":2089,"would_cite":true,"duration_ms":20035,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42B25","28A80"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims a complete, up-to-endpoints description of the weighted L^p power-weight range for spherical maximal operators, for every set of admissible radii, expressed through the Legendre–Assouad dimension function of the radius set","keywords":["power weight inequalities","spherical maximal function","restricted dilation sets","Legendre–Assouad function","Assouad spectrum","type set","weighted norm inequalities","fractal dimensions"],"falsifier":"Working directly in the proof, check the p=1 case of (5.9a,b) and the analogous derivative estimate (5.15b), which the paper says follow similarly. If either estimate carries an extra logarithmic factor in j, the j-sum in Proposition 5.1 would not close and the sufficiency direction would not hold as stated. A numerical or symbolic verification of these inequalities for d=2,3 would settle whether the argument is sound; alternatively, testing Lemma 4.1's power law on a set with strictly convex ν^♯ would test the sharpness of the lower boundary.","tokens_in":14109,"feed_emoji":"🌀","tokens_out":10022,"duration_ms":89945,"temperature":0.7,"pith_summary":"This paper claims a complete solution, up to endpoint cases, to an open problem in weighted harmonic analysis: decide exactly for which exponents p and which power weights |x|^α the spherical maximal operator M_E is bounded on L^p(|x|^α), for every nonempty set E of admissible radii. The answer is governed by two dimension-like invariants of E: the upper Minkowski-type exponent β and the Legendre–Assouad function ν^♯_E. The permitted region in the (1/p, α/p) plane lies between the straight line U(p) = (d−1)(p−1)−β and the curve L(p) = (d−1)(p−2)−(ν^♯)^†((d−1)(p−1)), with p ≥ 1 + β/(d−1). The result matters because previous work settled only special classes of dilation sets, whereas here the full range of p and α is described by a single geometric function of E.","feed_headline":"One fractal dimension curve now sets every weighted spherical bound","feed_subtitle":"A single Legendre–Assouad function fixes exactly which power weights work for any radius set.","key_machinery":"The central object is the Legendre–Assouad function ν^♯_E(ρ), a limsup of normalized entropy counts: for each scale δ, one maximizes |J|^{-ρ} N(E∩J,δ) over intervals J of logarithmic diameter between δ and 1, divides by log(δ^{-1}), and takes the limsup. It is convex, increasing, equals β for ρ ≤ 0 and equals ρ once ρ exceeds the quasi-Assouad dimension. Its generalized inverse (ν^♯)^† converts the covering geometry of E into the admissible range of α: the lower boundary of the type set is exactly where this inverse touches (d−1)(p−1). The paper also splits the spherical-average kernel into dyadic frequency pieces K^j_t whose Fourier decay rates carry the analytic part of the argument.","core_discovery":"Theorem 1.1 states that the closure of the type set of M_E on weighted L^p spaces is exactly the region {(1/p, α/p) : p ≥ p_β, L(p) ≤ α ≤ U(p)}, equivalently boundedness holds iff max{α+β, ν^♯((d−1)(p−2)−α)} ≤ (d−1)(p−1). The upper bound comes from a dyadic frequency decomposition of the spherical-average kernel and an entropy hypothesis; the lower bound comes from two families of test functions showing that M_E must be at least as large as the number of separated radii in small intervals dictates. Apart from endpoint cases in p and α, this settles the weighted boundedness question for arbitrary dilation sets.","pith_inferences":["Because the proof uses only the Fourier-decay and moment structure of the spherical-average kernel, the same Legendre–Assouad criterion is likely to govern other oscillatory convolution maximal operators with comparable decay; the paper does not pursue this.","The theorem identifies the Legendre–Assouad function as the sharp geometric statistic for weighted problems. A natural next step is to construct explicit compact radius sets realizing every admissible convex function ν^♯, which would show all shapes permitted by the theorem actually occur.","The proof leaves endpoints such as α = 1−d and p = p_β open; sharpening the interpolation step in Lemma 5.3 and the j-summation in Proposition 5.1 is the obvious route to a full boundary description.","One could use the same two test-function families with optimized parameters to probe related endpoint questions, such as restricted weak-type estimates at the boundary of the type set."],"forward_implications":["Every radius set now has a complete weighted boundedness description up to endpoints; the earlier open problem is resolved in full generality.","For the full set of all radii the formula recovers the classical range 1−d < α < (d−1)p−d, p > d/(d−1); for lacunary sets it recovers 1−d ≤ α < (d−1)(p−1), p > 1.","For Assouad-regular sets and finite unions of them, the type set is polygonal, with corners at scales determined by the quasi-Assouad dimensions of the components; for general sets it can be curved.","The entropy criterion is checkable: to decide whether a weighted inequality holds for a given E, it suffices to estimate covering numbers N(E∩J,δ) over logarithmic intervals, not to know the set's fine structure.","In the allowed range, M_E is bounded on the weighted space, so spherical means converge almost everywhere along E for functions in those weighted spaces."],"fun_headline_variants":["One fractal curve fixes all weighted spherical bounds","Legendre–Assouad settles weighted spherical maximal problem","Weighted spherical bounds: full range from one curve","Assouad spectrum pins down spherical maximal weights","Open problem closed: weighted L^p for spherical averages"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The sufficiency half rests on the dyadic kernel estimates (5.9a,b)–(5.11a,b): the frequency pieces of the spherical-average kernel must decay at the stated rates 2^{-j(d-1)/2} and 2^{-j(d-3)/2}, which in turn require the bump functions to have enough vanishing moments and the surface measure to decay like (1+|ξ|)^{-(d-1)/2}; the p=1 and derivative cases of these estimates are only sketched as analogous.","fun_headline_variants_meta":{"raw":{"variants":["One fractal curve fixes all weighted spherical bounds","Legendre–Assouad settles weighted spherical maximal problem","Weighted spherical bounds: full range from one curve","Assouad spectrum pins down spherical maximal weights","Open problem closed: weighted L^p for spherical averages"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00018,"raw_usage":{"total_tokens":1063,"prompt_tokens":592,"completion_tokens":471,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":336,"completion_tokens_details":{"reasoning_tokens":396}},"tokens_in":336,"tokens_out":471,"duration_ms":4305,"temperature":1.0,"reasoning_tokens":396,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T22:13:08.888035+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Working directly in the proof, check the p=1 case of (5.9a,b) and the analogous derivative estimate (5.15b), which the paper says follow similarly. If either estimate carries an extra logarithmic factor in j, the j-sum in Proposition 5.1 would not close and the sufficiency direction would not hold as stated. A numerical or symbolic verification of these inequalities for d=2,3 would settle whether the argument is sound; alternatively, testing Lemma 4.1's power law on a set with strictly convex ν^♯ would test the sharpness of the lower boundary.","supporting_citations":[],"review_version":1}