{"id":"442069d0-9b93-48dc-8e1d-0c4fe3079129","arxiv_id":"1908.05589","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"New multiscale polynomial Wolff axioms lead to Kakeya maximal estimates for p ≥ 1 + O(1/n), improving prior bounds in dimensions n=5 and n≥7.","lead":"This paper improves the best known bounds on the Kakeya maximal conjecture in five dimensions and in every dimension seven and higher. It also improves the Kakeya set conjecture for an infinite sequence of dimensions; the advance comes from a recursive polynomial partitioning argument and a new multiscale version of the polynomial Wolff axioms.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the main theorem rests on Lemma 3.5, whose induction is intricate but coherent; I found no concrete flaw that would invalidate Theorem 1.2.","rationale":"The reader correctly identifies Lemma 3.5 as the most novel and delicate ingredient, and I agree that it is the load-bearing point for the multiscale Wolff axioms. However, my reading did not surface a specific mathematical flaw: the induction over scales is internally consistent, the two lemmas use appropriate semialgebraic and polynomial tools, and the constants close in the claimed ranges. The nearest candidate issue is the 'if and only if' assertion (17), which is not literally true in general, but the later inclusion only needs the forward implication together with the section's projection property; the reverse direction is not used. The reader's conditional verdict is based on unproved side claims (Proposition 5.7 and Theorem 9.5), which are explicitly not needed for Theorem 1.2. Those unproved statements justify a request for proofs or references but do not change the status of the central theorem. A formal or computational re-verification of Lemma 3.5 would still be worthwhile, but I do not see grounds to move the verdict away from the reader's conditional acceptance.","tokens_in":38555,"tokens_out":46015,"duration_ms":453959,"concrete_test":"Verify the induction in Lemma 3.5 in a low-dimensional model: set n=4, k=2, m=3, take Z_2 and Z_3 to be nested coordinate planes, and compute |S_3(I_3,rho)| directly for a range of dyadic interval lengths lambda_2 <= lambda_3 and rho <= lambda_2. If the measured volume ever exceeds C rho^{-epsilon} (rho/lambda_2) lambda_3^3 rho, the multiscale volume bound fails. Separately, isolate (17) in a 2D model (lines through a point) to confirm the reverse implication is not used; the forward implication and projection equality should suffice for the covering in Lemma 3.7.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I read the proof as a chain whose key new link is the Wongkew-type bound Lemma 3.5. The two-step induction (Lemmas 3.6 and 3.7) is genuinely delicate, but the trigonometric covering argument and the algebraic average-value argument both close with the stated powers. The apparent 'iff' in (17) is an overstatement—the reverse direction is not needed; the covering uses only the forward direction plus the projection equality supplied by Corollary A.3—so it does not create a gap. The exponent bookkeeping in Section 8 checks out: conditions Xi=Yi=0 lead exactly to p = 1 + 2n/(n(n−1)+k(k−1)) and the Bourgain–Guth step to the min/max in Theorem 1.2. Proposition 5.7 and Theorem 9.5 are stated without proof but are explicitly non-load-bearing for Theorem 1.2. The proof is not machine-checked, and Lemma 3.5 is intricate enough that a formal re-verification would be valuable, but I cannot identify a concrete mathematical error.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves the Kakeya maximal conjecture in R^n for p at least 1 + min_{2<=k<=n} max{2n/((n-1)n+(k-1)k), 1/(n-k+1)} (Theorem 1.2), improving the previous best ranges in n=5 and all n>=7. The proof adapts Guth's polynomial-partitioning argument for the restriction problem to Kakeya, writing the induction as a recursive algorithm and introducing a multiscale version of the polynomial Wolff axioms (Theorem 1.4). The key new tool is a Wongkew-type volume bound for the semialgebraic sets S_m(I_m, rho) (Lemma 3.5), proved by a two-step induction mixing trigonometric and algebraic estimates. Theorem 1.2 follows from a k-broad estimate (Theorem 4.1) together with the Bourgain--Guth induction mechanism (Proposition 4.2). The paper also derives Hausdorff dimension bounds for Kakeya sets in an infinite sequence of dimensions (Corollary 9.1) and discusses variants such as the Guth--Zahl polynomial Wolff conjecture (Theorem 9.5).","tokens_in":38695,"tokens_out":11327,"duration_ms":105152,"significance":"If correct, this is a genuine advance on a central open problem in harmonic analysis, improving the Katz--Tao range in all sufficiently high dimensions under consideration and in dimension 5. The main technical novelty, Lemma 3.5, appears to be a substantial and plausible multiscale extension of Wongkew's theorem, and the overall proof is carefully structured: Theorem 1.2 is reduced to Theorem 4.1 and Proposition 4.2, Theorem 4.1 is proved via the recursive algorithms and the structural estimate in Section 8, and Theorem 1.4 is proved from Lemma 3.5. The argument is not circular: the target Kakeya maximal bounds are not assumed, and the proof builds on earlier single-scale polynomial Wolff axioms of Katz--Rogers and the Bourgain--Guth mechanism. The paper also gives explicit numerology, tables of exponents, and an honest discussion of the limitations and of the independent simultaneous work of Zahl (Remark 1.5). For these reasons the paper merits publication, provided the intricate proof of Lemma 3.5 is accepted as correct; I did not identify a concrete load-bearing error.","major_comments":[],"minor_comments":[{"comment":"The 'if and only if' assertion in (17) is stronger than what the subsequent argument needs; the proof only uses the implication from membership in L_ell(rho,t_ell) to containment of the line segment in S_ell(J,rho). The reverse implication should either be justified or the statement should be weakened accordingly.","section":"Section 3.2, Lemma 3.7"},{"comment":"The displayed inequality following the dyadic decomposition is easy to misread: the sum over J with the condition |S_ell(J,rho)| >= 4|J|rho^{n-1} appearing in the subscript should be typeset as a sum over {J in J : |S_ell(J,rho)| >= 4|J|rho^{n-1}} to avoid confusion with a product of factors.","section":"Section 3.2, proof of Lemma 3.5"},{"comment":"Proposition 5.7 is stated without proof. It is not used in the proof of Theorem 1.2, but as a proposition in the text it should either carry a proof or be explicitly labelled as a remark with an omitted proof.","section":"Section 5, Proposition 5.7"},{"comment":"Theorem 9.5 is presented as a theorem but no proof is supplied. Since it is not needed for the main Kakeya maximal theorem, this is not a blocking issue, but the text should clearly state that the proof is omitted or that the statement is a consequence of a variant of the earlier argument.","section":"Section 9.3, Theorem 9.5"},{"comment":"There are minor typographical errors: 'semiaglebraic' appears in the proof of Theorem 1.4 and 'algbraic' appears in the overview of Section 4; both should be corrected.","section":"Section 3 and Section 4"}],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know. First, this is a real advance: it is the first improvement over the Katz–Tao exponent in n=5 and all n≥7, and it comes with a new multiscale version of the polynomial Wolff axioms (Theorem 1.4) that is likely to be reused. Second, I read the proof looking for a reason to dismiss it, and I did not find one. The chain from Lemma 3.5 to Theorem 1.2 is long but coherent; the Wongkew-type volume bound is the delicate heart, and the two-step induction that proves it checks out. The recursive reformulation of Guth's induction is a genuinely new idea, not a cosmetic rewrite.\n\nThe paper is honest: Remark 1.5 acknowledges Zahl's independent simultaneous work on the same maximal estimates. That does not reduce the value here; the multiscale Wolff axioms are independently established and useful.\n\nThe soft spots are real but not load-bearing. Proposition 5.7 and Theorem 9.5 are stated without proof. They are side results and do not feed into Theorem 1.2, but in a paper this long the authors should either provide the arguments or give precise references. Also the proof is not machine-checked, so the usual medium risk of a hidden algebraic slip remains; the stress-test note raised a possible 'iff' overstatement in (17) and I agree it is only an overstatement, since the covering uses the forward direction plus Corollary A.3. No gap.\n\nCitation pattern is fine: the paper leans on Katz–Rogers and Hickman–Rogers for the single-scale case and builds outward; it does not assume the target result.\n\nBottom line: this is a serious paper for people working in harmonic analysis, geometric measure theory, and additive combinatorics. It deserves a careful referee, not a desk reject. I would bring it to the reading group and cite it. My recommendation is to accept with minor revision: fix the unproved side claims, either by proof or reference, and keep the presentation of Lemma 3.5 as is, maybe with a little more room for the induction.","headline":"Genuine first progress on the Kakeya maximal conjecture in n=5 and n≥7, with a credible new multiscale Wolff-axiom proof; worth serious refereeing.","tokens_in":39335,"tokens_out":2393,"would_cite":true,"duration_ms":24061,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42B25","28A78","14P10"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves the Kakeya maximal conjecture for all p ≥ 1 + min₂₌ₖ₌ₙ max{2n/((n−1)n+(k−1)k), 1/(n−k+1)}, improving the known range in dimensions 5 and all n ≥ 7.","keywords":["Kakeya maximal conjecture","Kakeya set conjecture","polynomial partitioning","multiscale polynomial Wolff axioms","broad norms","semialgebraic volume bounds","Hausdorff dimension","harmonic analysis"],"falsifier":"Compute $|S_m(I_m,\\rho)|$ for a small explicit configuration, for instance $n=4$, $m=3$, $k=2$, with $Z_2$ a plane, $Z_3$ a quadric surface, nested unit balls, and dyadic intervals $I_2, I_3$ chosen so that $\\rho/\\lambda_2$ and $\\rho/\\lambda_3$ are small but unequal, and compare the measured volume with the right-hand side of Lemma 3.5. Exceeding the claimed bound by more than the admissible $\\delta^{-\\varepsilon}$ factor would refute the lemma and the main theorem.","tokens_in":38265,"feed_emoji":"📐","tokens_out":11450,"duration_ms":99790,"temperature":0.7,"pith_summary":"This paper tries to establish that the Kakeya maximal conjecture holds for a wider range of Lebesgue exponents than was previously known, specifically for all $p \\geqslant 1 + \\min_{2 \\leq k \\leq n} \\max\\{2n/((n-1)n+(k-1)k), 1/(n-k+1)\\}$, improving the known range in five dimensions and in every dimension $n \\geqslant 7$. The proof adapts the polynomial partitioning method for Fourier restriction to the Kakeya problem by writing the induction argument as an explicit recursive algorithm, which brings out additional multiscale geometric information about direction-separated tubes. The authors show that such tubes satisfy a multiscale version of the polynomial Wolff axioms, with the proof resting on a new volume bound for certain semialgebraic sets formed by nested families of line segments. If the main estimate is correct, it also yields improved Hausdorff dimension lower bounds for Kakeya sets in an infinite sequence of dimensions, of the form $\\dim_H K \\geqslant (2-\\sqrt{2})n + 3/2 - 1/\\sqrt{2} - \\varepsilon$.","feed_headline":"Kakeya maximal bound widens in five dimensions and beyond","feed_subtitle":"A recursive polynomial-partitioning argument plus multiscale Wolff axioms widens the known range in n=5 and all n≥7.","key_machinery":"The load-bearing object is the multiscale polynomial Wolff axiom (Theorem 1.4): for direction-separated $\\delta$-tubes and nested varieties $Z_k,\\dots,Z_m$ of dimensions $k,\\dots,m$ with nested balls $B_{\\lambda_k} \\subseteq \\cdots \\subseteq B_{\\lambda_m}$, the number of tubes $T$ satisfying $|T \\cap B_{\\lambda_j} \\cap N_\\rho Z_j| \\geqslant \\lambda_j |T|$ for all $j = k,\\dots,m$ is bounded by $C_{n,d,\\varepsilon}(\\prod_{j=k}^{m-1} \\rho/\\lambda_j)(\\rho/\\lambda_m)^{n-m}\\delta^{-(n-1)-\\varepsilon}$. This estimate is sharp up to the $\\delta^{-\\varepsilon}$ factor, as nested planes show. The proof reduces to a new volume bound (Lemma 3.5) for the semialgebraic set $S_m(I_m,\\rho)$, the set of line segments that stay inside the $\\rho$-neighbourhood of all the varieties $Z_j$ over a common time interval $I_m$. That volume bound is proved by a two-step induction: a trigonometric estimate compares $S_{\\ell+1}$ with $S_\\ell$ at a slightly larger radius, and an algebraic estimate, built from polynomial parametrizations and degree-counting from algebraic geometry, governs how these sets expand over a longer dyadic interval.","core_discovery":"The central claim is Theorem 1.2: for every dimension $n \\geqslant 2$, the Kakeya maximal inequality $(K_p)$ holds for $p \\geqslant 1 + \\min_{2 \\leq k \\leq n} \\max\\{2n/((n-1)n + (k-1)k), 1/(n-k+1)\\}$. This yields the uniform bound $p \\geqslant 1 + (2-\\sqrt{2})^{-1}(n-1)^{-1}$ in all dimensions, which is strictly stronger than the previous best range, and gives new endpoints in low dimensions, such as $p = 18/13$ for $n=5$ and $p = 34/27$ for $n=7$. The proof proceeds by first proving a k-broad estimate (Theorem 4.1) for the Kakeya maximal function, then converting it to the usual linear maximal estimate via a broad-to-linear reduction. The broad estimate is obtained by feeding a recursive polynomial-partitioning algorithm into a new multiscale polynomial Wolff axiom (Theorem 1.4), which controls the number of tubes that have a substantial intersection with a whole nested chain of algebraic varieties at several scales.","pith_inferences":["Because Theorem 1.4 is sharp up to $\\delta^{-\\varepsilon}$ (nested planes saturate it), the multiscale volume bound is likely the true bottleneck for further endpoint improvements; if Lemma 3.5 could be upgraded to a logarithmic loss, the $\\varepsilon$ in Theorem 1.2 should disappear.","The recursive-algorithm formulation is essentially a way to optimize the trade-off between cellular and algebraic steps; one could program the algorithm to search for the best exponent range over all choices of the parameters $k, m, \\gamma_j$, potentially extending the method to other transversality problems.","The same $(2-\\sqrt{2})n$ asymptotics that appeared in the earlier sum-difference approach to Kakeya reappear here through a completely different route, suggesting that $2-\\sqrt{2}$ may be a structural constant for this family of polynomial partitioning arguments rather than an artifact of one technique.","The multiscale Wolff axioms imply k-linear estimates through the known dominance of k-linear over k-broad norms, so one might expect new multilinear restriction or Kakeya estimates at nearby exponents in dimensions covered by Theorem 1.4."],"forward_implications":["The Kakeya maximal conjecture is now known for $p \\geqslant 1 + (2-\\sqrt{2})^{-1}(n-1)^{-1}$ in every dimension, with strictly better exponents in dimensions 5 and all $n \\geqslant 7$.","Every Kakeya set in $\\mathbb{R}^n$ has Hausdorff dimension at least $(2-\\sqrt{2})n + 3/2 - 1/\\sqrt{2} - \\varepsilon$ for an infinite sequence of dimensions (Corollary 9.1).","The same multiscale Wolff axioms give an improved range for the polynomial Wolff axiom conjecture, namely $p \\geqslant 1 + \\min_{2 \\leq k \\leq n} \\max\\{(n/(n-1))^{n-k}, (n-1)/(n-k+1)\\}(n-1)^{-1}$ (Theorem 9.5).","The k-broad estimates hold for $p \\geqslant 1 + 2n/((n-1)n + (k-1)k)$, and each such estimate can be promoted to a genuine $L^p$ maximal inequality once $p$ is also at least $(n-k+2)/(n-k+1)$."],"supporting_citations":[{"why":"Supplies the single-scale polynomial Wolff axioms (Theorem 1.3) and the semialgebraic-section/Gromov-parametrization machinery that the multiscale Theorem 1.4 adapts.","marker":"[25]"},{"why":"Provides the polynomial partitioning theorem with cellular and algebraic cases and the k-broad norm framework used throughout the recursive algorithm.","marker":"[19]"},{"why":"Originates the adaptation of polynomial partitioning to restriction-type estimates that the paper transfers to the Kakeya maximal problem.","marker":"[18]"},{"why":"Supplies the classical volume bound for tubular neighbourhoods of algebraic varieties, the base ingredient that Lemma 3.5 generalises to nested multiscale semialgebraic sets.","marker":"[43]"},{"why":"Supplies the broad-to-linear mechanism (used as Proposition 4.2) that converts the k-broad maximal estimates into the stated linear Kakeya maximal inequality.","marker":"[7]"},{"why":"Provides the previous best range for the Kakeya maximal conjecture that Theorem 1.2 improves.","marker":"[27]"},{"why":"Confirms the polynomial Wolff axioms in four dimensions, forming a step toward the general Theorem 1.3 that the multiscale version extends.","marker":"[44]"},{"why":"Introduces the polynomial Wolff axiom conjecture and the four-dimensional estimates that Theorem 9.5 improves.","marker":"[22]"}],"fun_headline_variants":["New Kakeya bounds in 5D and all higher dimensions","Recursive algorithm sharpens Kakeya maximal estimates","Multiscale Wolff axioms yield new Kakeya bounds","Improved Kakeya maximal range: n=5 and n≥7","Kakeya maximal conjecture improves via multiscale geometry"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire argument rests on Lemma 3.5, the new volume bound for the semialgebraic sets $S_m(I_m,\\rho)$: if that bound fails at any of the nested scales and varieties produced by the recursive algorithm, then the multiscale polynomial Wolff axioms and hence the main maximal estimate no longer follow.","fun_headline_variants_meta":{"raw":{"variants":["New Kakeya bounds in 5D and all higher dimensions","Recursive algorithm sharpens Kakeya maximal estimates","Multiscale Wolff axioms yield new Kakeya bounds","Improved Kakeya maximal range: n=5 and n≥7","Kakeya maximal conjecture improves via multiscale geometry"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001166,"raw_usage":{"total_tokens":4800,"prompt_tokens":893,"completion_tokens":3907,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":509,"completion_tokens_details":{"reasoning_tokens":3822}},"tokens_in":509,"tokens_out":3907,"duration_ms":26655,"temperature":1.0,"reasoning_tokens":3822,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:24:55.385645+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $|S_m(I_m,\\rho)|$ for a small explicit configuration, for instance $n=4$, $m=3$, $k=2$, with $Z_2$ a plane, $Z_3$ a quadric surface, nested unit balls, and dyadic intervals $I_2, I_3$ chosen so that $\\rho/\\lambda_2$ and $\\rho/\\lambda_3$ are small but unequal, and compare the measured volume with the right-hand side of Lemma 3.5. Exceeding the claimed bound by more than the admissible $\\delta^{-\\varepsilon}$ factor would refute the lemma and the main theorem.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the single-scale polynomial Wolff axioms (Theorem 1.3) and the semialgebraic-section/Gromov-parametrization machinery that the multiscale Theorem 1.4 adapts."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Originates the adaptation of polynomial partitioning to restriction-type estimates that the paper transfers to the Kakeya maximal problem."},{"cited_title":"Wongkew, Volumes of tubular neighbourhoods of real algebraic variet ies, Pacific J","cited_arxiv_id":null,"evidence_quote":"Supplies the classical volume bound for tubular neighbourhoods of algebraic varieties, the base ingredient that Lemma 3.5 generalises to nested multiscale semialgebraic sets."},{"cited_title":"Bourgain and L","cited_arxiv_id":null,"evidence_quote":"Supplies the broad-to-linear mechanism (used as Proposition 4.2) that converts the k-broad maximal estimates into the stated linear Kakeya maximal inequality."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the previous best range for the Kakeya maximal conjecture that Theorem 1.2 improves."},{"cited_title":"Zahl, A discretized Severi-type theorem with applications to har monic analysis , Geom","cited_arxiv_id":null,"evidence_quote":"Confirms the polynomial Wolff axioms in four dimensions, forming a step toward the general Theorem 1.3 that the multiscale version extends."},{"cited_title":"Guth and J","cited_arxiv_id":null,"evidence_quote":"Introduces the polynomial Wolff axiom conjecture and the four-dimensional estimates that Theorem 9.5 improves."}],"review_version":1}