{"id":"f88587d0-19e3-478d-8205-4da25fd4d052","arxiv_id":"2509.06627","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For elliptic matrices with bounded measurable coefficients independent of the transversal variable, the parabolic Regularity problem is solvable in some L^p range, dual to the known Dirichlet range for the adjoint operator.","lead":"Parabolic partial differential equations with coefficients that do not vary in the direction transverse to the boundary are shown to have solvable L^p Regularity boundary value problems for all p in an interval (1,p0). The result closes a long-open question in harmonic analysis, completing the parabolic analogue of the elliptic theory for this class of coefficients.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.7's full interval (1,p0) is not covered by the proof if the [4] Dirichlet endpoint has p0'>2; Lemma 4.16 only establishes the required area-function bounds for 1<p≤2.","rationale":"The reader's conditional verdict identifies the same weakness: Lemma 4.16 only proves the area-function estimates for 1<p≤2, while Theorem 1.7 claims the full dual interval (1,p0). My stress-test confirms that this is the most load-bearing gap in the paper. The proof is otherwise coherent: the reduction in §3.2 is explicit, the kernel bounds and off-diagonal estimates are quoted from [4] with appropriate references, the Carleson-measure argument in §6 is self-contained, and the averaging estimates in §7 are detailed. No circularity or data fitting is present. The p-range issue is not merely cosmetic: block 1) of (3.16) is needed for the same exponent p as the final Regularity estimate, and the only available proof of that block stops at p=2. If the adjoint Dirichlet range from [4] happens to have endpoint p0'≥2, then p0≤2 and the stated theorem is fine; but the paper does not establish or cite this, and if p0'<2 the theorem as written goes beyond the proof. A straightforward weakening to p0=min(2,p0') would preserve a nontrivial result but would conflict with the claimed optimal dual range. Therefore the appropriate disposition remains CONDITIONAL, as the reader concluded; my read does not change that verdict.","tokens_in":34557,"tokens_out":7683,"duration_ms":68839,"concrete_test":"Locate in [4] (Auscher–Egert–Nyström) the exact exponent p0' such that the adjoint Dirichlet problem is solvable for all q>p0'. If p0'≥2, then p0≤2 and Lemma 4.16 suffices, so the concern is moot. If p0'<2, then p0>2 and the claimed interval includes p>2; to settle the concern, attempt to re-run the interpolation in §4.1 with an L^q endpoint q>2 in place of (4.1). If Proposition 4.3 cannot yield p>2 from the available H^1 and L^2 bounds, Theorem 1.7 must be restated with p0=min(2,p0') or with an additional p>2 area-function estimate.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is solvability of the Regularity problem for all 1<p<p0, with p0 dual to the adjoint Dirichlet endpoint p0' from [4]. The reduction in §3.2 assembles the proof from the four blocks listed in (3.16), and block 1) consists of five area-integral bounds. The only proof of block 1) is Lemma 4.16, whose statement is explicitly restricted to 1<p≤2 (Eq. 4.17). Its argument interpolates the L^2 bound (4.1) with the atomic L^1 bound (4.4) through Proposition 4.3; real interpolation between H^1_{1,1/2} and L^2_{1,1/2} cannot produce any p>2. Thus, if the Dirichlet endpoint from [4] satisfies p0'<2, equivalently p0>2, then Theorem 1.7 asserts Regularity solvability for exponents p∈(2,p0) for which the proof supplies no area-function control. The theorem statement does not impose p0≤2, and the 'moreover' clause presents the interval as the dual of the adjoint Dirichlet interval, so the reader is entitled to take p0 to be the actual dual endpoint from [4]. In that case the manuscript overclaims. A repair is available: one can replace p0 by min(2,p0') and correspondingly take the adjoint endpoint as max(2,p0'), but this weakens the claimed optimal dual range. Hence the concern is load-bearing for the theorem exactly as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the parabolic Regularity problem (R)_p for ∂_t u - div(A∇u) = 0 on Ω = O×R with bounded measurable uniformly elliptic coefficients independent of the transversal spatial variable x_{n+1}. Theorem 1.7 asserts that for some p0 > 1 the Regularity problem is solvable for all 1 < p < p0, and that this interval is dual to the range q > p0' on which the adjoint Dirichlet problem is solvable. The proof reduces (R)_p to the four blocks listed in (3.16): five area-function estimates, four nontangential maximal function estimates, a Carleson measure bound, and an area bound for the adjoint solution. The first block is proved for 1 < p ≤ 2 (Lemma 4.16), the second for 1 < p < ∞ (Lemma 5.9), the third for all p (Lemma 6.1), and the fourth in the Dirichlet-solvability range (Corollary 7.12).","tokens_in":34877,"tokens_out":4998,"duration_ms":43710,"significance":"If correct, this is a substantial result: it would complete the parabolic counterpart of the elliptic theory of Hofmann–Kenig–Mayboroda–Pipher for coefficients independent of the transversal variable, and it complements the Carleson-condition result of Dindoš–Li–Pipher. The paper's structure is a strength: the reduction in §3 is explicit, the estimates in §4–7 are proved in the paper rather than imported, and the dependence on the external results [4] and [40] is clearly identified. There are no fitted parameters and no circular arguments. The central caveat is the p-range gap in the area-function estimates, which affects Theorem 1.7 exactly as stated.","major_comments":[{"comment":"The area-function estimates that form block 1) of (3.16) are proved only for 1 < p ≤ 2. Lemma 4.16 is obtained by real interpolation (Proposition 4.3) between the L^2 bound (4.1) and the atomic L^1 bound (4.4), and real interpolation between an atomic H^1-type space and L^2 cannot produce exponents p > 2. These area-function estimates are used at the same exponent p throughout the reduction in §3.2, for example in the bounds for II_3, II_2, III_2, and V_1. If the adjoint Dirichlet endpoint supplied by [4] satisfies p0' < 2, then the dual exponent p0 = (p0')' exceeds 2, and Theorem 1.7 asserts solvability for p ∈ (2, p0) although the proof supplies no control of the corresponding area functions in that range. Theorem 1.7 presents (1, p0) as the dual interval without imposing p0 ≤ 2, so the statement overclaims. A repair is available by replacing p0 with min(2, p0') and correspondingly taking the adjoint endpoint as max(2, p0'), but this weakens the claimed optimal dual range. This is load-bearing for the theorem as stated.","section":"Lemma 4.16, Eq. (4.17); Theorem 1.7"},{"comment":"The abstract and the introductory paragraph of §1.1 describe the result as 'fully resolv[ing]' and 'optimally resolv[ing]' the range of solvability. In light of the gap described above, this optimality claim is not supported by the proof unless the area-function estimates are extended beyond p = 2 or the theorem is restated with the narrower interval. The authors should either prove the missing p > 2 estimates or explicitly qualify the optimality claim to the range established by Lemma 4.16.","section":"Abstract and §1.1"}],"minor_comments":[{"comment":"The integrand is written as |λ∇∥∂λP_{m,λ}f|^2 λ^{n-3} dxdt, but Definition 2.17 and the surrounding text require the power λ^{-n-3}; this appears to be a typographical error.","section":"§4.2, displayed formula after the sentence 'multiply both sides by λ^{-n-3}'"},{"comment":"The sentence 'By [4], this is true for all p′ > p0, where p0 > 1' overloads the symbol p0, which is also used in Theorem 1.7 for the Regularity endpoint; using q0 or p0' for the Dirichlet endpoint would avoid confusion.","section":"§3.1"},{"comment":"There are several typographical slips in the area-function section, including 'It them follows that Then' and the notation Γa_a for the away part of the cone; these should be corrected for readability.","section":"§4.1"},{"comment":"The expression C(λ E^{m-1}_λ ∂_j g)(y,s) is used as a pointwise quantity, whereas C(·) is defined as a supremum over boundary balls; the local Carleson expression and its L∞ norm should be distinguished notationally.","section":"§6, proof of Lemma 6.1"}],"recommendation":"major_revision","confidential_remarks":"The p-range gap in Theorem 1.7 is the only substantive mathematical obstacle I see. If the authors can extend the area-function estimates to p > 2, or alternatively restate the theorem and abstract with the provable interval, the paper would be suitable for publication. The rest of the proof appears coherent and the reduction is clearly executed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper resolves the last open case in the parabolic HKMP program: L^p Regularity solvability for operators whose coefficients are independent of the transversal variable x_{n+1}. That is a real result, and the proof is a serious piece of work. If it holds up, it completes the parabolic analogue of the elliptic story from [26] and answers a question that has been open for a while.\n\nWhat is genuinely new is the theorem itself and the technical machinery. The atomic Hardy-Sobolev interpolation in Section 4 and the time-averaging operator A_λ in Section 7 are new parabolic tools. The overall structure follows the elliptic roadmap, but there is real work in adapting it to the non-local half-derivative in time. There is no circularity; the paper leans on prior results [4], [12], [26] that have independent proofs, and I do not see fitted parameters anywhere.\n\nThe soft spot is the p-range. Lemma 4.16, the only proof of the area-function bounds in block 1) of (3.16), is limited to 1<p≤2. The argument interpolates between L^2 and atomic L^1, which cannot reach p>2. Theorem 1.7, however, states solvability for all 1<p<p0 where p0 is dual to the adjoint Dirichlet endpoint p0' from [4]. If p0'≥2, that is fine; but if p0'<2, then p0>2 and the interval (1,p0) overclaims—there is no area-function control for p∈(2,p0). The paper does not impose any condition like p0≤2, and the 'moreover' clause explicitly presents the interval as dual. So this is a load-bearing gap in the theorem exactly as stated.\n\nThe repair is straightforward: either quote the interval as (1,min(2,p0')) or prove the area bounds for p>2 using a different argument (perhaps duality from [11]). The rest of the proof—nontangential estimates, Carleson bound, the averaging operator—looks solid, though Section 7.1 is terse in places.\n\nWho should read this: anyone working on parabolic boundary value problems or on the elliptic-parabolic analogy. It deserves a serious referee; the issue I raised is fixable and not a sign of deeper trouble. I would send it out, but with an explicit request to address the range gap before acceptance.","headline":"Genuinely new result resolving the parabolic Regularity problem for transversally independent coefficients, but Theorem 1.7 overclaims the p-range: the area-function bounds are only proved for p≤2, so the full dual interval (1,p0) is not covered if p0>2.","tokens_in":35367,"tokens_out":3595,"would_cite":true,"duration_ms":31146,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35K20","35K10"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a uniformly elliptic, bounded measurable matrix independent of the transversal spatial variable, the parabolic $L^p$ Regularity problem is solvable on $(1,p_0)$ for some $p_0>1$, dual to the adjoint Dirichlet range.","keywords":["parabolic divergence-form operators","L^p regularity problem","transversally independent coefficients","nontangential maximal functions","square function estimates","Carleson measure","adjoint Dirichlet duality"],"falsifier":"A reader can look at the sharp Dirichlet exponent $r$ for a transversally independent operator in the paper's source [4]; if $r<2$, then $p_0=r/(r-1)>2$ and Theorem 1.7 asserts a range that the proof's Lemma 4.16 does not cover. In that case either an $L^p$ area-function estimate for $p>2$ must be proved or a counterexample at some $p\\in(2,p_0)$ would refute the theorem as stated.","tokens_in":34350,"feed_emoji":"🔥","tokens_out":12380,"duration_ms":106247,"temperature":0.7,"pith_summary":"This paper proves that the $L^p$ Regularity boundary value problem for the parabolic operator $\\partial_t u - \\operatorname{div}(A\\nabla u)=0$ on $\\Omega=O\\times\\mathbb{R}$ is solvable for a nontrivial interval of exponents $1<p<p_0$, $p_0>1$, whenever $A$ is uniformly elliptic, bounded and measurable, and its coefficients do not depend on the spatial variable $x_{n+1}$ transversal to the boundary. The Regularity problem is the harder companion to the Dirichlet problem: it demands nontangential convergence of the gradient to boundary data carrying one spatial derivative and a half time derivative in $L^p$. By duality, the interval $(1,p_0)$ is exactly the complement of the range $q>p_0'$ where the adjoint Dirichlet problem for $-\\partial_t u - \\operatorname{div}(A^*\\nabla u)=0$ is solvable. This closes the parabolic analogue of a gap that was already resolved in the elliptic setting for transversally independent coefficients.","feed_headline":"Parabolic Regularity solved for transversally independent coefficients","feed_subtitle":"The harder-than-Dirichlet regularity problem now has a dual-range L^p solution for this coefficient class.","key_machinery":"The central object is the resolvent family $P_\\lambda=(I+\\lambda^2 H_\\parallel)^{-m}$, where $\\lambda=x_{n+1}$ is the transversal variable and $H_\\parallel=\\partial_t - \\operatorname{div}_\\parallel(A_\\parallel\\nabla_\\parallel)$ is the tangential parabolic operator acting only on the boundary variables $(x,t)$. $P_\\lambda$ lifts boundary data $f$ into the interior, and all the terms in the boundary integral (3.10) are rewritten as evaluating square functions and nontangential maximal functions on combinations of $P_\\lambda f$, such as $A(\\partial_\\lambda P_\\lambda f)$, $A(\\lambda H_\\parallel P_\\lambda f)$, $A(\\lambda\\nabla_\\parallel\\partial_\\lambda P_\\lambda f)$, $A(\\lambda\\partial_\\lambda^2 P_\\lambda f)$ and $A(\\lambda^2\\nabla_\\parallel\\partial_\\lambda^2 P_\\lambda f)$. The machinery that makes the bounds work is the explicit kernel bound of [4], the identities $\\partial_\\lambda P_{\\lambda,m}=(2m/\\lambda)(P_{\\lambda,m+1}-P_{\\lambda,m})$ that reduce derivative terms to resolvent differences, the Caccioppoli inequality of Lemma 4.10, the atomic Hardy--Sobolev interpolation of [18] for $1<p\\le 2$, and the sharp maximal function for the nontangential estimates.","core_discovery":"Stated as Theorem 1.7, the paper's central claim is that for $\\Omega=O\\times\\mathbb{R}$ with $O$ an unbounded Lipschitz graph domain and $A$ satisfying (1.2) and (1.6), there exists $p_0>1$ such that the $L^p$ Regularity problem is solvable for all $1<p<p_0$, with the interval dual to the solvability range of the adjoint Dirichlet problem. The proof reduces this to eleven explicit $L^p$ bounds: five square-function bounds, four nontangential maximal-function bounds, a Carleson measure bound, and one commutator estimate, all applied to the family $P_\\lambda f=(I+\\lambda^2 H_\\parallel)^{-m}f$ with $\\lambda=x_{n+1}$. Sections 4--7 establish these bounds: area functions by atomic Hardy--Sobolev interpolation for $1<p\\le 2$, nontangential maximal functions by sharp-maximal-function estimates for $1<p<\\infty$, the Carleson bound by induction on the resolvent power, and the final commutator term through the averaging operator $A_\\lambda$ and a time-averaging lemma. The proof for a general Lipschitz graph domain follows from the half-space case by a change of variables that preserves the transversal independence condition.","pith_inferences":["The stated interval $(1,p_0)$ should be read against the proof's range: Lemma 4.16 proves the area-function bounds only for $1<p\\le 2$, so the theorem is fully established for intervals with $p_0\\le 2$; if the adjoint Dirichlet exponent inherited from [4] is below 2, the interval with $p_0>2$ would require extending Lemma 4.16.","The same resolvent-and-duality scheme is a natural template for the parabolic Neumann problem, since in the elliptic case Regularity solvability has often been the stepping stone to Neumann solvability; that connection is not made in this paper.","A testable extension would be to perturb the transversal independence by a coefficient whose gradient satisfies a small Carleson measure condition, asking whether the dual-range $(1,p_0)$ persists; the companion Carleson-condition paper [12] solves a different range, so a synthesis is not immediate."],"forward_implications":["If the theorem is right, the parabolic Regularity problem with data in $\\dot L^p_{1,1/2}(\\partial\\Omega)$ is solvable for every $1<p<p_0$, so gradients of energy solutions converge nontangentially to data with one spatial and a half time derivative.","Theorem 2.22 then gives the additional $L^p$ control of the half-time derivative and the time-adapted second-order term, so this solution concept coincides with the classical Regularity formulation used for the heat equation.","The dual-range statement means that any improvement in the adjoint Dirichlet range for transversally independent coefficients translates directly into a wider $(1,p_0)$ interval for Regularity.","The change of variables extends the result to all unbounded Lipschitz graph domains $O\\times\\mathbb{R}$, not only the half-space, without losing transversal independence.","The eleven-bound reduction is a reusable checklist that could be applied to related parabolic boundary value problems such as Neumann problems."],"supporting_citations":[{"why":"supplies the adjoint Dirichlet solvability, the L2 square-function bounds, and the kernel and Carleson estimates used throughout","marker":"[4]"},{"why":"provides the elliptic Regularity roadmap, including the boundary-integral reduction and the Fubini exchange for the V-term","marker":"[26]"},{"why":"introduces the adjoint inhomogeneous Dirichlet problem and the energy-class framework the parabolic proof adapts","marker":"[12]"},{"why":"establishes the L2 well-posedness and energy spaces that define the solution class","marker":"[3]"},{"why":"gives the atomic Hardy--Sobolev spaces and the interpolation result that extends bounds down to 1<p<=2","marker":"[18]"},{"why":"provides Theorem 2.22, showing spatial-gradient control implies the half-time-derivative bounds in the Regularity estimate","marker":"[10]"},{"why":"supplies the weighted parabolic Kato perspective and kernel estimates the authors describe as methodologically relevant","marker":"[1]"},{"why":"gives the Lp bound for the area term A(lambda h) used in Corollary 7.10","marker":"[40]"}],"fun_headline_variants":["Parabolic Regularity solved for transversal-independent coefficients","Lp Regularity proven: parabolic operators with transversal independence","Dual-range Lp solvability for parabolic Regularity problem","Parabolic Regularity: beyond Dirichlet, now resolved","Transversal independence yields Lp Regularity for parabolic PDEs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the exponent $p_0$ in Theorem 1.7 can be taken at most $2$, because Lemma 4.16 proves the necessary area-function estimates only for $1<p\\le 2$ and no extension to $p>2$ is supplied.","fun_headline_variants_meta":{"raw":{"variants":["Parabolic Regularity solved for transversal-independent coefficients","Lp Regularity proven: parabolic operators with transversal independence","Dual-range Lp solvability for parabolic Regularity problem","Parabolic Regularity: beyond Dirichlet, now resolved","Transversal independence yields Lp Regularity for parabolic PDEs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000899,"raw_usage":{"total_tokens":3934,"prompt_tokens":1071,"completion_tokens":2863,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":687,"completion_tokens_details":{"reasoning_tokens":2781}},"tokens_in":687,"tokens_out":2863,"duration_ms":17884,"temperature":1.0,"reasoning_tokens":2781,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:17:03.736982+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A reader can look at the sharp Dirichlet exponent $r$ for a transversally independent operator in the paper's source [4]; if $r<2$, then $p_0=r/(r-1)>2$ and Theorem 1.7 asserts a range that the proof's Lemma 4.16 does not cover. In that case either an $L^p$ area-function estimate for $p>2$ must be proved or a counterexample at some $p\\in(2,p_0)$ would refute the theorem as stated.","supporting_citations":[{"cited_title":"Auscher, M","cited_arxiv_id":null,"evidence_quote":"supplies the adjoint Dirichlet solvability, the L2 square-function bounds, and the kernel and Carleson estimates used throughout"},{"cited_title":"Hofmann,C","cited_arxiv_id":null,"evidence_quote":"provides the elliptic Regularity roadmap, including the boundary-integral reduction and the Fubini exchange for the V-term"},{"cited_title":"Auscher, M","cited_arxiv_id":null,"evidence_quote":"establishes the L2 well-posedness and energy spaces that define the solution class"},{"cited_title":"Dindoˇ s,On the regularity problem for parabolic operators and the role of half time derivative, J","cited_arxiv_id":null,"evidence_quote":"provides Theorem 2.22, showing spatial-gradient control implies the half-time-derivative bounds in the Regularity estimate"},{"cited_title":"Egert, K","cited_arxiv_id":null,"evidence_quote":"supplies the weighted parabolic Kato perspective and kernel estimates the authors describe as methodologically relevant"},{"cited_title":"Ulmer,L p Boundary Value Problems for Elliptic and Parabolic Operators, University of Edinburgh, thesis (2024)","cited_arxiv_id":null,"evidence_quote":"gives the Lp bound for the area term A(lambda h) used in Corollary 7.10"}],"review_version":2}