{"id":"e67f3e1c-ca59-4178-b98b-3d7b17270218","arxiv_id":"1909.00447","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Dirichlet problem for complex k-Hessian equations on compact Hermitian manifolds with boundary is solved under the assumption that a smooth subsolution exists.","lead":"This paper proves that k-Hessian equations, a family of nonlinear equations from complex geometry, can be solved on compact curved spaces with boundary whenever a smooth subsolution exists. The key is a new boundary estimate whose scaling lets the authors remove the most difficult gradient obstruction.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The blow-up step in Section 6 needs a detailed weak-convergence verification for the rescaled k-Hessian equations; without it the Liouville-based gradient estimate is not fully established.","rationale":"A close reading of Sections 2–6 shows that the internal estimate chain — C^0 bound, boundary gradient, mixed normal-tangential estimate of order K^{1/2}, double-normal boundary estimate, and the interior Hessian bound of Proposition 3.3 — is coherent. The apparent issue in Section 5 about the eigenvalues of W lying in a bounded set is resolved once one reads w as built from the subsolution rather than the solution; the OCR ambiguity between the underlined subsolution and the solution caused the impression of a gap. The only truly load-bearing point that is not written out is the passage to the limit in the blow-up argument of Section 6: the rescaled k-Hessian equations converge to the homogeneous equation, after which the Liouville theorem applies. This is a standard but nontrivial verification, particularly with varying Hermitian backgrounds, and it is exactly the condition needed for the gradient estimate. The reader's weakest_assumption identifies the same point, so I agree with the reader's assessment. Since the paper's overall argument is coherent and the cited convergence is plausible, the verdict ACCEPT remains appropriate; the concern is a verification step rather than a demonstrated flaw.","tokens_in":27962,"tokens_out":54923,"duration_ms":495939,"concrete_test":"Write out the weak-limit passage in detail. For each compact K ⊂ C^n and each continuous f, show ∫_K f (χ_i + √−1∂∂û_i)^k ∧ α_i^{n−k} → ∫_K f (√−1∂∂u∞)^k ∧ β^{n−k}, using the C^{1,γ} convergence of û_i, the uniform L∞ bound on ∂∂û_i, and the explicit forms χ_i = M_i^{−2}χ and α_i = M_i^{−2}α. In particular, verify that the error terms involving χ_i and α_i−β vanish as M_i→∞ and that the Bedford–Taylor/Blocki product is continuous under this convergence. If this verification succeeds, the blow-up argument stands; if it fails, Proposition 6.1 and hence Theorem 1.1 lack a proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Proposition 6.1 is the load-bearing step for the central claim. To prove the gradient estimate, the paper rescales a sequence ui by Mi = sup|∇ui| and obtains a C^{1,γ} limit u∞ with |∇u∞|(0)=1. The key assertion is that the rescaled equations (χ_i + √−1∂∂û_i)^k ∧ α_i^{n−k} = ψ_i α_i^n, with χ_i = M_i^{−2}χ and α_i = M_i^{−2}α, pass to the limit (√−1∂∂u∞)^k ∧ β^{n−k} = 0 in the Bedford–Taylor/Blocki sense (Section 6, after (6.3)). The contradiction then invokes the Dinew–Kołodziej Liouville theorem [19]. This weak-convergence passage is cited (e.g., Demailly [17], Dinew–Kołodziej [19]) but not written out. If the k-Hessian measures failed to converge — for instance because û_i are only C^{1,γ} and the Hermitian backgrounds vary, or because the product is not continuous along this particular rescaled family — the gradient bound would not follow, and Theorem 1.1 would lack a proof. The rest of the a priori program (C^0 bounds, boundary mixed normal-tangential estimate of order K^{1/2}, boundary double-normal estimate, and Proposition 3.3) is internally coherent; this is the single point where the argument depends on an external convergence theorem in a nontrivial way.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves the Dirichlet problem for the k-Hessian equation on a compact Hermitian manifold with boundary, assuming the existence of an admissible smooth subsolution. The main theorem (Theorem 1.1) states that, under this hypothesis, there is a unique smooth admissible solution with prescribed boundary data. The proof is the standard a priori-estimate route: the continuity method reduces the problem to uniform C^0, C^1, and C^2 bounds (Theorem 2.1); the C^0 and tangential boundary gradient bounds are obtained by the comparison principle (Lemmas 3.1–3.2); the interior second-order bound is reduced to a boundary bound via the Hou-Ma-Wu/Székelyhidi maximum principle (Proposition 3.3); the boundary second-order bound is proved in Sections 4–5 through a K^{1/2} mixed normal-tangential estimate (Proposition 4.1) and a double-normal estimate using a Caffarelli-Nirenberg-Spruck barrier (Theorem 5.1, Proposition 5.2), yielding sup |√-1∂∂u| ≤ C(1+sup |∇u|^2); and the gradient bound is obtained in Section 6 by a blow-up argument that derives a bounded nonconstant solution of the homogeneous complex k-Hessian equation on C^n, contradicting the Liouville theorem of Dinew-Kołodziej (Proposition 6.1). The logic of the proof is coherent and the scaling of each estimate is carefully matched to the blow-up argument.","tokens_in":28268,"tokens_out":32547,"duration_ms":313325,"significance":"If the result is correct, this is a significant contribution to complex analysis and PDE on Hermitian manifolds. It solves the global Dirichlet problem for complex k-Hessian equations for all 1 ≤ k ≤ n, extending the Guan-Li solution for Monge-Ampère equations and improving on prior work of Gu-Nguyen and Feng-Ge-Zheng, which required additional hypotheses (locally conformally Kähler assumptions or gradient estimates via maximum principle). The main novelty is the boundary second-order estimate with the sharp K^{1/2} scale for the mixed normal-tangential derivatives, which is precisely what makes the Liouville-based blow-up argument work; this avoids the long-standing open problem of a maximum-principle gradient estimate. The paper is clearly written, the constants are tracked, and the argument is self-contained up to standard external theorems (Gårding's inequality, the Hou-Ma-Wu/Székelyhidi interior estimate, and the Dinew-Kołodziej Liouville theorem). The proof of the gradient estimate via contradiction and rescaling is an elegant application of existing weak-compactness tools for complex Hessian operators.","major_comments":[],"minor_comments":[{"comment":"The passage from the rescaled equations to the limiting homogeneous equation (√-1∂∂u∞)^k ∧ β^{n-k}=0 is highly compressed. Since this is the critical step in which the Dinew-Kołodziej Liouville theorem is applied, I recommend adding a short justification: after multiplying the equation by M_i^{2(n-k)}, the rescaled Hermitian metrics M_i^2 f^*α converge smoothly to the Euclidean form β, the rescaled χ-terms tend to zero, and the weak continuity of the complex Hessian operator for locally uniformly convergent sequences (Blocki [3], Demailly [17]) then gives the claimed limit. This would remove any doubt about the hypotheses of the convergence theorem.","section":"Section 6, after (6.3)"},{"comment":"The same symbol û_i is used for the rescaled solution and later for the rescaled subsolution, which makes (6.8) confusing; I suggest using distinct notation such as û_i for the solution and ̲u_i (or a different letter) for the subsolution.","section":"Section 6, Case 2b"},{"comment":"The author names 'Blocki' and 'Kołodziej' appear garbled as 'B/suppress locki' and 'Ko/suppress lodziej' in the extracted text; these should be corrected.","section":"Section 1 and references"},{"comment":"The phrase 'ψ /greaterorequalslantc > 0' should read 'ψ ≥ c > 0'.","section":"Section 2.1 and abstract"},{"comment":"The sentence 'One easily checks that the sequences û_i and b̂_i converge in C^{1,γ/2} on compact sets of {Im z_n > 0} ∪ {0} to constant functions u∞ = φ(p∞) = b∞' would benefit from a brief explanation that this follows from the smoothness of u and b and the fact that the rescaled arguments tend to p∞ uniformly on compact sets.","section":"Section 6, Case 2b"}],"recommendation":"minor_revision","confidential_remarks":"The stress-test concern about the weak convergence in Section 6 is, in my reading, adequately addressed by the citations to Blocki and Dinew-Kołodziej; the passage is terse but standard, and the required normalization is implicit. I do not see a fatal gap. The paper is a strong contribution and suitable for publication after minor revisions addressing the clarity points listed in the report."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me give you my read. The paper proves the Dirichlet problem for the k-Hessian equation on compact Hermitian manifolds with boundary, for 1<k<n, assuming a subsolution. That was open; prior work covered only locally conformally Kähler metrics or small balls, or needed extra positivity. The main new ingredient is a boundary second-order estimate of the form |normal-tangential| ≤ C K^{1/2} and |double-normal| ≤ C K, with K = 1 + sup|∇u|^2. That scaling is exactly what a blow-up argument needs. The proof builds on the Caffarelli-Nirenberg-Spruck barrier and the Hou-Ma-Wu/Székelyhidi interior estimate, and it is worked out in detail for the complex Hermitian setting.\n\nThe paper is honest about what is new: the K^{1/2} mixed estimate and the blow-up gradient estimate are not in the cited literature. The continuity method reduction is standard, and the a priori estimate program is coherent. I did not find circularity; the quoted Liouville theorem, the Gårding inequality, and the interior estimates are independent results.\n\nSoft spots: Section 6's passage from the rescaled equations to the limiting homogeneous equation (√−1∂∂u∞)^k ∧ β^{n−k} = 0 is compressed. It cites Demailly and Dinew–Kołodziej, but does not spell out the normalization or the weak convergence of the k-Hessian measures. For an expert this is standard, but it is the one load-bearing step that relies on an external theorem in a non-obvious way. The transition from Theorem 5.1 to Proposition 5.2 also has a typo: the proof of Theorem 5.1 starts by invoking Proposition 3.3, which already contains the boundary term; the intended meaning is clear, but the text is muddled. These are repairable and do not threaten the main theorem.\n\nThe paper deserves a serious referee. It settles a recognized open problem, the method is reusable, and the technical core is sound. I would send it out and expect the referee to ask for a longer write-up of Section 6 and a cleanup of Section 5. Citation pattern looks appropriate. I'd bring it to reading group, and would cite it once the final version is out.","headline":"Solves the Dirichlet problem for complex k-Hessian equations on Hermitian manifolds with boundary; the K^{1/2} boundary estimate is genuinely new and the blow-up argument works, though Section 6 compresses a standard but nontrivial weak-convergence step.","tokens_in":100,"tokens_out":9058,"would_cite":true,"duration_ms":93657,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35J60","32W20","53C55"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the Dirichlet problem for complex k-Hessian equations on compact Hermitian manifolds with boundary is solvable whenever a smooth subsolution exists, and that the solution is unique.","keywords":["k-Hessian equation","Dirichlet problem","complex manifold","Hermitian manifold","a priori estimates","blow-up argument","Liouville theorem","subsolution"],"falsifier":"Find a bounded, non-constant entire function $u$ on $\\mathbb{C}^n$ satisfying $(\\sqrt{-1}\\partial\\bar{\\partial} u)^k \\wedge \\beta^{n-k} = 0$ in the weak sense, or construct a sequence of solutions on a fixed manifold with boundary for which $\\sup_X |\\sqrt{-1}\\partial\\bar{\\partial} u|$ grows faster than $C(1 + \\sup_X |\\nabla u|^2)$. Either would break the central estimate.","tokens_in":27736,"feed_emoji":"📐","tokens_out":5140,"duration_ms":554806,"temperature":0.7,"pith_summary":"This paper proves that the Dirichlet problem for complex k-Hessian equations on a compact Hermitian manifold with boundary has a unique smooth admissible solution, provided a smooth subsolution exists. The interest is that this is the first global result for 1 < k < n on manifolds with boundary, where the usual maximum-principle gradient estimate is an open problem. The authors obtain the required a priori bounds by a new boundary second-order estimate whose scale is the gradient norm squared, and then apply a blow-up argument with a Liouville theorem. A sympathetic reader should take away that the missing piece was not ellipticity or subsolution theory but the correct scaling of the boundary second-order estimate.","feed_headline":"k-Hessian boundary problem solved on compact complex manifolds","feed_subtitle":"A boundary estimate scaled by the gradient squared, plus a blow-up Liouville argument, yields smooth solutions from a subsolution.","key_machinery":"The load-bearing device is the boundary second-order estimate at gradient scale: for solutions on a manifold with boundary, the mixed normal-tangential second derivatives satisfy $|h_{\\bar{n} i}|(0) \\leq C K^{1/2}$, and the double-normal derivative satisfies $h_{\\bar{n} n} \\leq C K$, with $K = 1 + \\sup_X |\\nabla u|^2_{X,\\alpha}$. These are obtained by barrier constructions of B. Guan and of Caffarelli-Nirenberg-Spruck, using the elementary symmetric polynomials $\\sigma_k$ and the Gårding cone. Combined with the Hou-Ma-Wu interior estimate, they give $\\sup_X |\\sqrt{-1}\\partial\\bar{\\partial} u|_{X,\\alpha} \\leq C K$. The scale $K^{1/2}$ is then matched by a blow-up argument: if $|\\nabla u|$ were unbounded, rescaling would produce a bounded entire solution of the homogeneous equation $(\\sqrt{-1}\\partial\\bar{\\partial} u)^k \\wedge \\beta^{n-k} = 0$, contradicting the Liouville theorem of Dinew-Kołodziej.","core_discovery":"The central claim is Theorem 1.1: given a compact Hermitian manifold $(X,\\alpha)$ with boundary, a $k$-positive $(1,1)$-form $\\chi$, a positive function $\\psi$, boundary data $\\phi$, and a smooth subsolution $\\underline{u}$ with $\\sigma_k(\\lambda(\\underline{u})) \\geq \\psi$ and $\\underline{u}|_{\\partial X} = \\phi$, there is a unique smooth $u$ solving $\\sigma_k(\\lambda(u)) = \\psi$ with $u|_{\\partial X} = \\phi$ and $\\lambda(u) \\in \\Gamma_k$. The proof reduces this to a priori estimates: $C^0$, $C^1$, and $C^2$ bounds. The new ingredient is a boundary second-order estimate of the mixed normal-tangential and double-normal derivatives at scale $K^{1/2}$, where $K = 1 + \\sup_X |\\nabla u|^2_{X,\\alpha}$. This scale is what allows a blow-up argument, using the Liouville theorem for the homogeneous $k$-Hessian equation, to close the gradient estimate; the interior $C^2$ estimate of Hou-Ma-Wu then yields the full bound.","pith_inferences":["The same blow-up-with-Liouville strategy might yield gradient estimates for other fully nonlinear equations (for example, Lagrangian phase or Hessian quotient equations) on manifolds with boundary, whenever a Liouville theorem is available for the rescaled equation.","The boundary estimate may be sharp: if the mixed normal-tangential estimate could not be improved below $K^{1/2}$, the blow-up argument would fail, so the scale is likely forced by the structure of $\\sigma_k$.","A testable extension would be to adapt the argument to parabolic k-Hessian flows with boundary data; the same scaling should give long-time existence and convergence under a subsolution condition."],"forward_implications":["The Dirichlet problem for k-Hessian equations is solvable on all compact Hermitian manifolds with boundary that admit a subsolution, not only on domains in $\\mathbb{C}^n$ or under curvature assumptions on the boundary.","The result extends the complex Monge-Ampère Dirichlet theory (the case $k=n$) to all $1 \\leq k \\leq n$ in the same subsolution framework.","The scale $K^{1/2}$ in the boundary estimate gives the quantitative control needed for blow-up arguments, suggesting a template for other fully nonlinear equations on manifolds with boundary.","Along the continuity path, the a priori bounds give uniform ellipticity and hence $C^{2,\\alpha}$ and higher regularity of solutions."],"supporting_citations":[{"why":"Supplies the double-normal barrier technique and the $\\Gamma_k$ cone properties used in Section 5.","marker":"[8]"},{"why":"Gives the interior second-order estimate for complex Hessian equations used in Proposition 3.3.","marker":"[43]"},{"why":"Extends the interior second-order estimate to Hermitian manifolds, used in Section 3.","marker":"[67]"},{"why":"Provides the Liouville theorem for bounded entire solutions of the homogeneous k-Hessian equation, the contradiction in the blow-up step.","marker":"[19]"},{"why":"Introduces the subsolution approach and the barrier $v = (u-\\underline{u}) + c_0 d - N d^2$ used for the boundary estimate.","marker":"[31]"},{"why":"The analogous blow-up argument for the complex Monge-Ampère equation with boundary, which the present paper generalizes.","marker":"[4]"},{"why":"A related boundary estimate for a more general class of equations, which the authors improve to the $K^{1/2}$ scale.","marker":"[24]"},{"why":"Solves the complex Monge-Ampère Dirichlet problem on manifolds with a subsolution, the $k = n$ case.","marker":"[33]"},{"why":"Provides the weak/pluripotential framework used in taking the limit of the rescaled equations in the blow-up step.","marker":"[3]"}],"fun_headline_variants":["Gradient-scaled boundary estimate solves k-Hessian Dirichlet problem","Subsolution suffices for k-Hessian Dirichlet problem on complex manifolds","Blow-up argument with gradient scale solves k-Hessian Dirichlet problem","Boundary estimate yields k-Hessian Dirichlet solution","Dirichlet problem for k-Hessian solved via subsolution and boundary estimates"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on the hypothesis that a smooth subsolution exists, and within the argument the gradient bound stands on the Liouville theorem that bounded entire solutions of the homogeneous k-Hessian equation are constant; if either fails, the conclusion does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Gradient-scaled boundary estimate solves k-Hessian Dirichlet problem","Subsolution suffices for k-Hessian Dirichlet problem on complex manifolds","Blow-up argument with gradient scale solves k-Hessian Dirichlet problem","Boundary estimate yields k-Hessian Dirichlet solution","Dirichlet problem for k-Hessian solved via subsolution and boundary estimates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001021,"raw_usage":{"total_tokens":4259,"prompt_tokens":850,"completion_tokens":3409,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":466,"completion_tokens_details":{"reasoning_tokens":3311}},"tokens_in":466,"tokens_out":3409,"duration_ms":25388,"temperature":1.0,"reasoning_tokens":3311,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:53:19.411936+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a bounded, non-constant entire function $u$ on $\\mathbb{C}^n$ satisfying $(\\sqrt{-1}\\partial\\bar{\\partial} u)^k \\wedge \\beta^{n-k} = 0$ in the weak sense, or construct a sequence of solutions on a fixed manifold with boundary for which $\\sup_X |\\sqrt{-1}\\partial\\bar{\\partial} u|$ grows faster than $C(1 + \\sup_X |\\nabla u|^2)$. Either would break the central estimate.","supporting_citations":[{"cited_title":"Caffarelli, L","cited_arxiv_id":null,"evidence_quote":"Supplies the double-normal barrier technique and the $\\Gamma_k$ cone properties used in Section 5."},{"cited_title":"Hou, X.-N","cited_arxiv_id":null,"evidence_quote":"Gives the interior second-order estimate for complex Hessian equations used in Proposition 3.3."},{"cited_title":"Sz´ ekelyhidi,Fully-nonlinear elliptic equations on compact Hermitian m anifolds, J","cited_arxiv_id":null,"evidence_quote":"Extends the interior second-order estimate to Hermitian manifolds, used in Section 3."},{"cited_title":"Dinew and S","cited_arxiv_id":null,"evidence_quote":"Provides the Liouville theorem for bounded entire solutions of the homogeneous k-Hessian equation, the contradiction in the blow-up step."},{"cited_title":"Guan, The Dirichlet problem for complex MongeAmpere equations an d regularity of the pluri-complex Green function , Comm","cited_arxiv_id":null,"evidence_quote":"Introduces the subsolution approach and the barrier $v = (u-\\underline{u}) + c_0 d - N d^2$ used for the boundary estimate."},{"cited_title":"Boucksom, Monge-Ampere equations on complex manifolds with boundary , in Com- plex Monge-Ampere Equations and Geodesics in the Space of Ka hler Metrics, V","cited_arxiv_id":null,"evidence_quote":"The analogous blow-up argument for the complex Monge-Ampère equation with boundary, which the present paper generalizes."},{"cited_title":"Guan and Q","cited_arxiv_id":null,"evidence_quote":"Solves the complex Monge-Ampère Dirichlet problem on manifolds with a subsolution, the $k = n$ case."},{"cited_title":"Blocki, Weak solutions to the complex Hessian equation , Ann","cited_arxiv_id":null,"evidence_quote":"Provides the weak/pluripotential framework used in taking the limit of the rescaled equations in the blow-up step."}],"review_version":1}