{"id":"f5419d50-d779-4cef-9317-7480ec76f23a","arxiv_id":"2502.02964","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Weak solutions of the m-polyharmonic Dirichlet problem in Reifenberg-flat domains are C^{m-1,α} up to the boundary, with an a priori estimate in terms of the L^q norm of the source.","lead":"This paper proves that solutions of higher-order elliptic equations with zero boundary data are Hölder continuous up to the boundary even when the domain is only Reifenberg-flat, a class that includes fractal boundaries. The proof extends the Nirenberg translation method to arbitrary operator order and is a substantive advance in boundary regularity theory, though it contains a gap in a critical case.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Critical case 2m=N: Proposition 5.2's exponent condition (5.18) is impossible for q=2, leaving Theorem 1.1 unproved in that case; repairable by a sharper estimate.","rationale":"I read the paper in good faith. The overall strategy is coherent: flat-boundary translation regularity (Theorem 4.1), a compactness transfer to Reifenberg-flat boundaries (Proposition 5.1), a replacement-argument energy decay (Proposition 5.2), and finally Campanato embedding (Theorem 6.1). The reader's CONDITIONAL verdict is appropriate. I focused on the critical case 2m=N because it is explicitly included in Theorem 1.1 and the written exponent condition in Proposition 5.2 is impossible for q=2: solving q(2−p′)≥2p′ gives p′≤2q/(q+2), which for q=2 means p′≤1, whereas every finite Sobolev conjugate has p′>1. This is a concrete algebraic gap in the proof of the main theorem, not just a self-containment issue. I did not select the dependence on the authors' prior paper [12] as the single concern: that reliance is real but the cited result is published and the required stable-domain facts are standard in this context; it affects self-containedness more than likely correctness. I also noticed that the normalization in Proposition 5.1 appears to use r_n^{N/2−2m} where scaling suggests r_n^{N/2−m}; I treat that as a typographical slip rather than a load-bearing flaw. The critical-case gap is repairable by a sharper exponent estimate that keeps the kη−N and η terms, so the appropriate disposition remains CONDITIONAL rather than REJECT or ACCEPT.","tokens_in":27427,"tokens_out":22762,"duration_ms":218696,"concrete_test":"Re-derive the exponent comparison in Proposition 5.2 for 2m=N, q=2, writing p′=1+δ. Test whether γ_k := 2Nk(1/(1+δ) − 1/2) satisfies γ_k ≥ (k+1)(N−η) for all k≥k0 when δ<η/(2N) and k0 ≥ max(N/η, (N−η)/(η−2Nδ)). If yes, replace condition (5.18) with this choice of δ and k0; if no such δ exists, Proposition 5.2 fails in the critical case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The weakest load-bearing point is the exponent bookkeeping in Proposition 5.2 in the critical case 2m=N. The proof needs, for the second term in (5.16), the inequality γ_k := k·2N(q−p′)/(p′q) ≥ (k+1)(N−η). The paper reduces this to condition (5.18), q(2−p′) ≥ 2p′, and then claims it can be met for 2m=N by taking p′ arbitrarily close to 1. But solving (5.18) gives p′ ≤ 2q/(q+2); for q=2 this forces p′ ≤ 1, impossible for the conjugate of any finite Lebesgue exponent, since p′ > 1. The estimate (5.17) also discards the positive-looking terms kη−N and η too early. As written, therefore, the proof of Proposition 5.2 does not establish the energy decay in the critical case, and since the statement of Theorem 1.1 includes q=2 when 2m=N, Theorem 6.1 is not fully proved. This is a genuine proof gap rather than a disproof: the inequality can likely be rescued by keeping the kη−N+η terms and choosing p′=1+δ with δ<η/(2N) and k0 sufficiently large.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript establishes C^{m-1,α} regularity up to the boundary for weak solutions u∈H^m_0(Ω) of the constant-coefficient elliptic Dirichlet problem A(u)=f in Reifenberg-flat domains Ω, with f∈L^q under the scaling condition mq≥N if 2m<N and q=2 otherwise. The proof combines a Nirenberg translation method adapted to operators of arbitrary order (interior and flat-boundary versions, Theorems 3.1 and 4.1), a compactness argument that transfers the flat-boundary decay to Reifenberg-flat boundaries (Proposition 5.1), a polyharmonic replacement argument for nonzero right-hand sides (Proposition 5.2), and Campanato embedding to conclude the Hölder estimate (Theorem 6.1). The main theorem is Theorem 1.1/6.1.","tokens_in":27683,"tokens_out":14803,"duration_ms":135136,"significance":"If completed, the result is significant: it gives C^{m-1,α} boundary regularity for higher-order elliptic equations on domains whose boundary need only be Reifenberg-flat, hence possibly only C^{0,γ} or fractal. The extension of the Nirenberg translation method to arbitrary order is a genuine technical contribution, and the replacement/decay framework is well suited to the higher-order Dirichlet problem. The writing is generally clear and the argument is convincing outside the critical case. However, one load-bearing exponent estimate in Proposition 5.2 is not proved for 2m=N, and the proof relies on imported stable-domain facts from [12] that should be stated precisely.","major_comments":[{"comment":"The exponent bookkeeping in the critical case 2m=N is incorrect. Solving q(2−p′)≥2p′ for p′ gives p′≤2q/(q+2), not p′≤2q/(1+q) as stated in Case 2. For q=2 the correct bound is p′≤1, which is incompatible with the value p′=1+δ that is available when 2m=N. Consequently the displayed lower bound γ−(k+1)(N−η)≥η in (5.17) is not justified in this case, and the proof of the energy decay (H_k) does not go through for q=2, which is precisely the case allowed by Theorem 1.1 and Theorem 6.1. The gap appears repairable: instead of discarding the nonnegative terms kη−N and η, one can keep them and choose p′=1+δ with δ<η/(2N) and k0≥N/η, which makes k(η−2Nδ)−N+η positive for all k≥k0. The same repair is needed in Proposition 5.3, whose proof is declared to be identical.","section":"Proposition 5.2, Eqs. (5.17)-(5.18), Case 2"},{"comment":"The proof of Proposition 5.1 uses two facts imported from the authors' previous paper [12]: the characterization H^m_0(Ω)={u∈H^m(R^N): u=0 a.e. on R^N\\Ω} for Reifenberg-flat domains, and the approximation of a function in H^m(B(0,1)) vanishing on B(0,1)∩{x_N<0} by functions compactly supported in B+(0,1). These facts are load-bearing for Claim A3 and for the zero-extension used throughout Section 5. Since the introduction describes the paper as \"completely self-contained\", the precise statement of the imported theorem and a verification that its hypotheses cover (ε0,r0)-Reifenberg-flat domains should be added. If [12] indeed supplies these facts, this is a clarification issue rather than a mathematical error; without them, the contradiction argument in Proposition 5.1 is incomplete.","section":"Section 2, Eq. (2.1); Section 5, Claim A3"}],"minor_comments":[{"comment":"Proposition 5.1 is applied to v_k, but v_k is only defined on B_k∩Ω and equals u on the artificial boundary; strictly speaking the proposition requires a function in H^m_0(Ω). The application is justified if v_k is extended by u outside B_k, since the extension lies in H^m_0(Ω) and is A-harmonic in B_k∩Ω; please state this extension explicitly.","section":"Section 5, after Eq. (5.8)"},{"comment":"The constants in the statement do not match the proof: the proof yields ∫_{Q_r} v² ≤ 4/(1−λ)∫_{Q^λ_r} v² + 8r²∫_{Q_r}(∂_N v)², whereas (2.8) is written with coefficients 4 and 3r. This appears to be a typo, but it should be corrected because the displayed inequality is used in the proof of Theorem 4.1.","section":"Lemma 2.2, Eq. (2.8)"},{"comment":"The formula α=(λ−N)/p appears to contain a typo; from the definition of the Campanato seminorm the correct exponent is α=(λ−N)/2. The subsequent use in Proposition 2.4 is consistent with the corrected formula.","section":"Proposition 2.3"},{"comment":"The symbol C_A is used both for the ellipticity constant of A and, implicitly, for the constant coming from Proposition 5.1 in (5.8)-(5.10). Please use distinct notation (for example C_1 for the decay constant) throughout the replacement argument.","section":"Proposition 5.2, notation"}],"recommendation":"major_revision","confidential_remarks":"The exponent error in Proposition 5.2 is localized and the proposed repair is plausible, so I do not recommend rejection. The reliance on [12] is acceptable, but the paper should not be described as completely self-contained without stating the imported theorem. The novelty and scope appear adequate for the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on arXiv:2502.02964. The paper proves boundary regularity up to C^{m-1,α} for the m-polyharmonic Dirichlet problem in Reifenberg-flat domains, for arbitrary m. That's new. The main tools are a Nirenberg translation method adapted to high-order operators, an induction on the operator order, and a compactness argument using polyharmonic replacements. The interior and flat-boundary regularity sections (Theorems 3.1 and 4.1) are detailed and self-contained, and the paper is honest about where the difficulty lies: you can't just differentiate in the normal direction.\n\nThe main theorem is plausible and, if the proof gap gets fixed, this will be a solid paper. But as written, there is a real gap in the proof of Proposition 5.2, in the critical case 2m = N. The authors need the inequality q(2−p′) ≥ 2p′ to hold with p′ close to 1 when q=2. Solving that inequality gives p′ ≤ 2q/(q+2), and for q=2 this forces p′ ≤ 1, which is impossible for any Lebesgue conjugate. The paper instead uses p′ ≤ 2q/(1+q), which is weaker and doesn't make (5.18) true. So the energy decay in that case is not established. This is not a counterexample—the inequality can likely be rescued by keeping the positive terms kη−N and η, choosing p′=1+δ with δ small, and taking k0 large. But as written, Theorem 1.1 is not fully proved.\n\nA smaller point: the paper claims to be \"completely self-contained\" but relies on the authors' previous paper [12] for the stable-domain/Mosco convergence results, including the characterization (2.1) that is used throughout. That's not a fatal flaw—it's a published result and self-citation is fine when the cited work is real—but the self-contained claim is overstated.\n\nMy overall assessment: the result is significant, the method is new, and the paper is mostly carefully written. The critical-case gap is the one thing that needs to be fixed before this should be published. I'd send it to peer review, with instructions to focus on Proposition 5.2 and the dependence on [12]. If the authors can close that gap, I'd be happy to cite it.","headline":"Strong result, likely correct, but Proposition 5.2 has a real gap in the critical case 2m=N that needs repair before Theorem 1.1 is established.","tokens_in":28248,"tokens_out":3261,"would_cite":false,"duration_ms":30601,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B65","35J40","35D30"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that weak solutions of $m$-th order elliptic Dirichlet problems are $C^{m-1,\\alpha}$ up to the boundary on Reifenberg-flat domains, even when the boundary may be fractal.","keywords":["polyharmonic operators","Reifenberg-flat domains","boundary regularity","translation method","Hölder continuity","energy decay","Dirichlet problem"],"falsifier":"Consider the planar sector $\\Omega_\\varepsilon=\\{(r,\\theta): 0<\\theta<\\pi+\\varepsilon\\}$ with the clamped condition for $m=2$, and compute the leading exponent $\\lambda(\\varepsilon)$ in the boundary expansion of the solution to $\\Delta^2 u=1$ with $u=\\partial_\\nu u=0$ on the two sides. The sector is Reifenberg-flat with flatness of order $\\varepsilon$, so Theorem 1.1 predicts that for every $\\alpha<1$, one can make $\\lambda(\\varepsilon)>1+\\alpha$ by taking $\\varepsilon$ small. If a numerical or asymptotic computation of the characteristic equation for $\\lambda(\\varepsilon)$ showed that $\\lambda(\\varepsilon_j)\\le 1+\\alpha$ for some fixed $\\alpha>0$ along a sequence $\\varepsilon_j\\to0$, then the energy-decay claim behind the theorem would be false.","tokens_in":27190,"feed_emoji":"📐","tokens_out":22641,"duration_ms":235593,"temperature":0.7,"pith_summary":"This paper proves that a weak solution of an $m$-th order elliptic Dirichlet problem is H\\\"older continuous up to the boundary, together with its derivatives through order $m-1$, on domains that are only Reifenberg-flat: rough sets whose boundary lies within a small fraction of the radius of a hyperplane at every point and scale, and which may even be fractal. The result applies to any constant-coefficient symmetric elliptic operator of order $2m$, including the polyharmonic operator $(-\\Delta)^m$, and gives an a priori bound by the $L^q$ norm of the right-hand side. What makes this surprising is that the boundary itself is less regular than Lipschitz; the extra smoothness comes from the homogeneous Dirichlet condition, which forces the solution and its derivatives up to order $m-1$ to vanish at the boundary. If the theorem is correct, it completes the boundary regularity picture for the polyharmonic Dirichlet problem in a broad class of rough domains and shows a loss of only one derivative compared with interior regularity.","feed_headline":"Polyharmonic solutions stay Hölder-smooth up to rough flat boundaries","feed_subtitle":"Weak solutions gain regularity up to order m−1 even when the boundary is fractal but close to a plane.","key_machinery":"The load-bearing object is the boundary energy-decay estimate $\\int_{B(x,r)}|\\nabla^m u|^2\\,dx \\le C (r/R)^{N-b}\\int_{B(x,R)}|\\nabla^m u|^2\\,dx$ for boundary points $x$, combined with a standard integral-decay-to-H\\\"older criterion that converts such decay into $C^{m-1,\\alpha}$ regularity. The paper's main technical novelty is a self-contained proof of the flat-boundary case by the translation method: horizontal difference quotients are admissible test functions, and one then recovers vertical derivatives by an induction on the order using the decomposition $A = B - \\partial_N D(\\partial_N u)$, where $D$ is an elliptic operator of order $2m-2$. A key lemma (Lemma 2.1) guarantees the elliptic factor $D$, and a one-dimensional integral inequality bridges the one-derivative gap that would otherwise block the induction. The transfer from a hyperplane to a Reifenberg-flat domain is made by a compactness argument whose strong-convergence step uses a convergence notion for the spaces $H^m_0$ under Hausdorff convergence, inherited from the authors' earlier work on stable domains.","core_discovery":"The central claim is Theorem 1.1: for every $\\alpha\\in(0,1)$, if $q\\ge 2$ with $m q\\ge N$ when $2m<N$ (and $q=2$ otherwise), there exist flatness and scale parameters $\\varepsilon_0>0$ and $r_0>0$ such that for every $(\\varepsilon_0,r_0)$-Reifenberg-flat domain $\\Omega$ and every $f\\in L^q(\\Omega)$, the weak solution $u\\in H^m_0(\\Omega)$ of $A(u)=f$ belongs to $C^{m-1,\\alpha}(\\Omega)$ and satisfies $\\|u\\|_{C^{m-1,\\alpha}(\\Omega)}\\le C\\|f\\|_{L^q(\\Omega)}$. This extends the known $m=1$ Poisson result to all orders $m$, and it is sharp: near conical boundary points one can construct solutions that are $C^{m-1,\\alpha}$ but no better, so the boundary roughness exactly costs one derivative at the top order. The proof locates the mechanism in an energy-decay estimate at boundary points, a quantified statement that $\\nabla^m u$ has no mass accumulating at the boundary faster than a power of the radius.","pith_inferences":["The energy-decay mechanism suggests the same $C^{m-1,\\alpha}$ boundary regularity should persist for divergence-form operators with H\\\"older continuous coefficients, as the authors note; the compactness argument would then need to be re-run with a coefficient-freezing step.","The technique might extend to quasilinear higher-order problems wherever a Caccioppoli-type energy decay can be proved, with the flat-boundary translation method being the main obstacle to overcome.","A testable numerical consequence: on Reifenberg-flat domains, finite element convergence rates limited by boundary roughness should improve by exactly $m-1$ orders compared with generic rough-boundary PDEs, because the Dirichlet condition forces derivatives up to order $m-1$ to vanish at the boundary.","The paper's reliance on the stable-domain convergence theory suggests that if that convergence fails for a subfamily of Reifenberg-flat domains, the regularity might still hold but would require a different mechanism; the threshold $\\varepsilon_0$ may be linked to the constant in the stable-domain theory."],"forward_implications":["For the standard Laplacian ($m=1$), the theorem recovers and extends the previously known Poisson regularity result on Reifenberg-flat domains, without needing an a priori $L^p$ bound on $u$.","For the biharmonic operator ($m=2$) in dimension $3$, any $H^2_0$ solution of $\\Delta^2 u=f$ with $f\\in L^2$ is $C^{1,\\alpha}$ up to the boundary for every $\\alpha<1$ on Reifenberg-flat domains.","The boundary regularity of order $m-1$ is sharp: singular examples at conical boundaries show one cannot expect $C^{m,\\alpha}$ regularity of the highest derivatives, and the homogeneous Dirichlet condition is what purchases the extra regularity.","The proof supplies a self-contained $H^{2m}$ up-to-the-boundary regularity theory for constant-coefficient elliptic operators of order $2m$ in smooth domains, via the translation method adapted to arbitrary order.","For Reifenberg-flat domains with fractal boundary, the result gives uniform a priori estimates depending only on flatness and scale constants, not on the boundary's Hausdorff dimension."],"supporting_citations":[{"why":"Supplies the stable-domain identity $H^m_0(\\Omega)=\\{u\\in H^m(\\mathbb R^N): u=0 \\text{ a.e. on }\\Omega^c\\}$ and the convergence of $H^m_0$ under Hausdorff convergence used in the compactness step of Proposition 5.1.","marker":"[12]"},{"why":"The prior $m=1$ Poisson regularity result on Reifenberg-flat domains that this work extends and compares against, including the sharp cone example.","marker":"[18]"},{"why":"Supplies the integral-decay-to-H\\\"older regularity criterion that converts the energy decay estimate into $C^{m-1,\\alpha}$ boundary regularity.","marker":"[11]"}],"fun_headline_variants":["Polyharmonic weak solutions gain Hölder regularity on rough flat domains","Rough flat boundaries don't break polyharmonic Hölder smoothness","Polyharmonic regularity on Reifenberg-flat domains up to order m-1","Hölder smoothness for polyharmonic solutions near flat rough boundaries","One derivative loss: polyharmonic solutions stay Hölder on flat rough domains"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on the authors' earlier stable-domain theory: for Reifenberg-flat domains, $H^m_0(\\Omega)$ is exactly the set of $H^m$ functions vanishing outside $\\Omega$, and these spaces converge in a variational sense when the domains converge in Hausdorff distance; if that theory fails for some Reifenberg-flat domain, the compactness step that transfers flat-boundary decay to rough boundaries fails and the proof collapses.","fun_headline_variants_meta":{"raw":{"variants":["Polyharmonic weak solutions gain Hölder regularity on rough flat domains","Rough flat boundaries don't break polyharmonic Hölder smoothness","Polyharmonic regularity on Reifenberg-flat domains up to order m-1","Hölder smoothness for polyharmonic solutions near flat rough boundaries","One derivative loss: polyharmonic solutions stay Hölder on flat rough domains"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000497,"raw_usage":{"total_tokens":2397,"prompt_tokens":866,"completion_tokens":1531,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":482,"completion_tokens_details":{"reasoning_tokens":1451}},"tokens_in":482,"tokens_out":1531,"duration_ms":11849,"temperature":1.0,"reasoning_tokens":1451,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T10:25:01.052980+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Consider the planar sector $\\Omega_\\varepsilon=\\{(r,\\theta): 0<\\theta<\\pi+\\varepsilon\\}$ with the clamped condition for $m=2$, and compute the leading exponent $\\lambda(\\varepsilon)$ in the boundary expansion of the solution to $\\Delta^2 u=1$ with $u=\\partial_\\nu u=0$ on the two sides. The sector is Reifenberg-flat with flatness of order $\\varepsilon$, so Theorem 1.1 predicts that for every $\\alpha<1$, one can make $\\lambda(\\varepsilon)>1+\\alpha$ by taking $\\varepsilon$ small. If a numerical or asymptotic computation of the characteristic equation for $\\lambda(\\varepsilon)$ showed that $\\lambda(\\varepsilon_j)\\le 1+\\alpha$ for some fixed $\\alpha>0$ along a sequence $\\varepsilon_j\\to0$, then the energy-decay claim behind the theorem would be false.","supporting_citations":[{"cited_title":"Stable domains for higher order elliptic operators","cited_arxiv_id":null,"evidence_quote":"Supplies the stable-domain identity $H^m_0(\\Omega)=\\{u\\in H^m(\\mathbb R^N): u=0 \\text{ a.e. on }\\Omega^c\\}$ and the convergence of $H^m_0$ under Hausdorff convergence used in the compactness step of Proposition 5.1."},{"cited_title":"Boundary regularit y for the Poisson equation in Reifenberg-flat domains","cited_arxiv_id":null,"evidence_quote":"The prior $m=1$ Poisson regularity result on Reifenberg-flat domains that this work extends and compares against, including the sharp cone example."},{"cited_title":"Introduction to regularity theory for nonlinear elliptic sy stems","cited_arxiv_id":null,"evidence_quote":"Supplies the integral-decay-to-H\\\"older regularity criterion that converts the energy decay estimate into $C^{m-1,\\alpha}$ boundary regularity."}],"review_version":1}