{"id":"fbd13bc7-e9ac-451f-aa3b-216549c4ec41","arxiv_id":"2505.24188","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Conformally compact Lovelock metrics are polyhomogeneous near the boundary, and the leading obstruction to smoothness in even dimensions is the Einstein ambient obstruction tensor scaled by a coupling constant.","lead":"This paper proves that curved spaces satisfying Lovelock equations, a family of higher-curvature generalizations of Einstein's equation, have well-behaved boundary expansions of the kind used in conformal geometry and the AdS/CFT correspondence. It also computes the leading boundary obstruction to smoothness and gives topological obstructions to filling a boundary by such spaces.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 6.1 is stated without the nondegeneracy condition A1(α)≠0, but the proof in §6.2 explicitly requires it; for Lovelock couplings with A1=0 the linearized gauge-fixed operator degenerates.","rationale":"The reader's weakest_assumption correctly identifies the same gap: Theorem 1.1 and Theorem 6.1 are stated without the nondegeneracy condition A1(α)≠0, while the proof explicitly invokes this condition to justify the ellipticity of the gauge-fixed linearization and the independence of the recursive equations (6.6)–(6.7). My reading of §4, §6.2 and §6.3 confirms that A1(α) is not merely a convenient constant but the leading coefficient of the entire linearized operator; if it vanishes, every step from Lemma 6.2 through Theorem 6.9 loses its ellipticity. The concern is therefore load-bearing and not a matter of stylistic preference. I agree with the reader's CONDITIONAL verdict rather than escalating to REJECT, because the proof appears sound for the generic case A1(α)≠0, and the statement can likely be repaired by adding the hypothesis to the theorems (or by a separate argument for the degenerate case). The proposed concrete test with α1=72, α2=1 in n=6 isolates exactly the degeneracy in a minimal example and would settle whether the stated theorem is false or merely missing an argument.","tokens_in":37708,"tokens_out":2396,"duration_ms":30574,"concrete_test":"No additional test beyond the one specified above is needed; the n=6, α1=72, α2=1 computation is the decisive check.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that every asymptotically hyperbolic Lovelock metric is polyhomogeneous (Theorem 6.1, and the equivalent Theorem 1.1). The proof, however, only works under the hypothesis 1 ∈ LimSec(α), which by the definition of LimSec(α) includes A1(α) ≠ 0. In §6.2 the text states: 'The assumption on LimSec(α) implies that A1(α) is not equal to 0,' and this nonvanishing is then used to make equations (6.6) and (6.7) linearly independent, solving inductively for the expansion coefficients. The same coefficient A1(α) multiplies the entire Laplace-type linearization in Lemma 6.2 and Lemma 6.5; when A1(α)=0 the modified Lovelock tensor Qα(·, t) is not elliptic at the hyperbolic metric, so the elliptic-regularity engine of §5 and §6 collapses. The stated theorems do not exclude this case. For example, with n=6 and α1=72, α2=1 (using the paper's normalization) one obtains A1(α)=72−72=0 while the hyperbolic metric still satisfies the corresponding Lovelock equation. Thus the proof as written does not cover a nonempty, explicitly definable class of asymptotically hyperbolic Lovelock metrics; either the theorem must be restricted to A1(α)≠0 or the degenerate case must be handled separately.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies conformally compact (asymptotically hyperbolic) metrics satisfying the Lovelock equations, a higher-curvature generalization of the Einstein equation. Its main claim is that every asymptotically hyperbolic Lovelock metric is polyhomogeneous near the conformal boundary, thereby realizing formally the Fefferman--Graham expansion obtained by Albin. It also computes the leading-order ambient obstruction tensor in even dimensions, constructs formal solutions of the singular Yamabe-(2q) problem, and proves an index-theoretic obstruction to conformally compact Lovelock fillings under a scalar-curvature lower bound. The proof strategy adapts Biquard--Herzlich's elliptic-regularity approach from the Einstein case, using DeTurck gauge fixing, indicial analysis, and Green's integral operators on a hyperbolic semi-ball.","tokens_in":37991,"tokens_out":7942,"duration_ms":97231,"significance":"If the main regularity theorem holds, it is a substantial extension of the polyhomogeneity theory of conformally compact Einstein metrics to a large class of Lovelock theories, and it gives a rigorous justification for treating Albin's formal expansions as actual asymptotic expansions. The obstruction-tensor computation in Section 7 and the index obstruction in Section 9 are concrete and potentially useful for constructing examples. The paper is not merely programmatic: the algebraic computations of the Lovelock linearization and the obstruction tensor are explicit, and the proof follows the cited framework of BH14 in detail. However, the central theorem as stated is stronger than the proof supports because a nondegeneracy hypothesis is silently assumed.","major_comments":[{"comment":"The statement of Theorem 6.1 omits the hypothesis 1 ∈ LimSec(α) (equivalently A1(α) ≠ 0) that the proof requires. In the proof of Theorem 6.6, the text states 'The assumption on LimSec(α) implies that A1(α) is not equal to 0' and uses this to make equations (6.6) and (6.7) linearly independent, allowing the inductive determination of the expansion coefficients. The same nonvanishing is needed for the linearization formula in Lemma 6.2 and Lemma 6.5, where the displayed operator is proportional to A1(α); if A1(α) = 0, the linearized modified Lovelock tensor degenerates at the hyperbolic metric and the elliptic-regularity engine of Section 6 collapses. The excluded case is nonempty: for n = 6, the choice α1 = 72, α2 = 1 gives A1(α) = 0, while the hyperbolic metric still satisfies the Lovelock equation by the normalization of λ(2q). The theorem should either be restricted to nondegenerate couplings A1(α) ≠ 0, or the degenerate case must be analyzed separately.","section":"§6.2, Theorem 6.1 (and Theorem 1.1)"},{"comment":"There is an inconsistency in the definition of A1(α) between the statement of Theorem 7.2 (and Theorem 1.2) and the definition used in Section 4 and Section 6.2. In §4 and §6.2, A1(α) is defined as ∑_q α_q (−1/2)^{q−1} ((n−2)!/2) (2q)!/(n−2q)!, while Theorem 7.2 states A1(α) = ∑_q α_q (−1/2)^{q−1} ((n−2)!/2) (2q−1)!/(n−2q)!. These differ by a factor of 2q. Since the leading obstruction term in formula (7.1) is A1(α)/c_n times a generically nonzero curvature expression, the two versions yield different invariants and cannot both be correct. Please reconcile the definitions and verify the displayed formula with the correct constant.","section":"§7, Theorem 7.2; compare §4 and §6.2"},{"comment":"The proof of the singular Yamabe-(2q) result is substantially sketched. In particular, the sentence introducing 'A key observation based on the commutativity of Cg and ∂x' is not a proof; the extraction of ∂x^{s+1}u from the term C_g^{2q−1}(T^{q−1}·η) is the main technical step, and it is asserted rather than demonstrated. Since Theorem 8.1 is one of the stated results of the paper, the derivation of the recurrence (8.6) and the subsequent handling of the log term should be written out in sufficient detail to be checked.","section":"§8, proof of Theorem 8.1"}],"minor_comments":[{"comment":"In the definition of the Kulkarni-Nomizu product for simple double forms, the displayed formula 'ω.η = (α1 ∧ β1) ⊗ (α1 ∧ β2)' appears to contain a typo; it should presumably be '(α1 ∧ β1) ⊗ (α2 ∧ β2)'.","section":"§3.2 (Kulkarni-Nomizu product)"},{"comment":"The sentence 'As discussed in Example 5.6, the Laplace term of Lϕ0 is an isomorphism' suppresses the coefficient A1(α) in front of both Laplace-type factors; the isomorphism statement is only valid when A1(α) ≠ 0, so the role of the nondegeneracy hypothesis should be made explicit in the displayed argument.","section":"§6.2, proof of Lemma 6.2"},{"comment":"The transition from the displayed derivative formula to 'This leads to the same result for the highest-order term in the derivatives of h' is terse: the reader is asked to compare with [GH05, Theorem 2.1] without a full explanation of why the 'terms involving ∂x^k h with k < s' do not contribute to the trace-free leading-order term. A short justification or a more detailed computation would improve readability.","section":"§7, proof of Theorem 7.2"},{"comment":"In the final paragraph of Section 8, the notation alternates between 'β_q' and 'βg' in the expression for eFβ(g); please use a single consistent notation for the components of β.","section":"§8, notation"},{"comment":"The proof of Lemma 9.7 says that the argument of [CQY04, Lemma 2.1] 'only uses the Fefferman-Graham expansion and thus holds' for Lovelock metrics. This is plausible, but since the Lovelock expansion can contain logarithmic terms in even boundary dimensions, the statement that the eigenfunction has the displayed form without a log term should be checked explicitly or referenced to the relevant part of the paper.","section":"§9.2, Lemma 9.7"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's main defect is a missing hypothesis in Theorems 1.1 and 6.1, not an error in the overall strategy; the proof is a faithful adaptation of BH14 and the algebraic sections are mostly explicit. I would not reject on this basis, but the statements need to be made consistent with the assumptions used in the proofs. I also recommend asking the author to resolve the conflicting definitions of A1(α) before publication, since a referee cannot determine the correct coefficient of the obstruction tensor from the manuscript as it stands."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Core take: this is a genuine extension of Biquard–Herzlich to Lovelock metrics, but the main theorem is stated too broadly. Theorem 6.1 (and Theorem 1.1) claim polyhomogeneity for every asymptotically hyperbolic Lovelock metric; the proof in §6.2 uses the nondegeneracy condition 1 ∈ LimSec(α), which implies A1(α) ≠ 0. That condition is needed for the linearization of the modified Lovelock operator to be elliptic and for the induction at order n−1. The stress-test example n=6, α1=72, α2=1 gives A1=0 with the hyperbolic metric still a solution, so the degenerate case is nonempty. The proof does not reach it. This is fixable — state the theorem with the LimSec/A1 condition, or treat the degenerate case separately — but as written the headline overclaims.\n\nWhat's genuinely new: the polyhomogeneity argument for Lovelock metrics, the ambient obstruction computation in Section 7, and the singular Yamabe-(2q) recurrence. The paper follows BH14 carefully and is transparent about using Albin's linearization formulas; the computations in Section 7 are explicit and checkable. No circularity, and the citation pattern is honest.\n\nThe other soft spots are Sections 8 and 9. The singular Yamabe theorem constructs a formal expansion but does not prove existence of an actual function with that expansion; and the CCL filling obstruction relies on [CQY04] and [GHS21] in ways that are plausible but not fully proved. Both are likely repairable but are not complete as written.\n\nWho gets value: people working on conformal geometry, AdS/CFT, and higher-curvature gravity, especially those using the Fefferman–Graham expansion. The paper should go to a serious referee. I recommend sending it out with the expectation that the main theorem be restated with the nondegeneracy hypothesis and Sections 8–9 be expanded.","headline":"Real extension of BH14 to Lovelock metrics, but Theorem 1.1 is stated without the A1 nonzero hypothesis the proof actually needs; fixing that should be the referee's first demand.","tokens_in":38554,"tokens_out":4970,"would_cite":true,"duration_ms":62039,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C18","53C21","53C25","58J20"],"pacs":[],"model":"deepseek-v4-flash","headline":"Conformally compact Lovelock metrics admit polyhomogeneous boundary expansions, generalizing the Einstein asymptotics.","keywords":["conformally compact metrics","Lovelock tensors","polyhomogeneous expansions","Fefferman-Graham expansion","ambient obstruction tensor","singular Yamabe problem","Dirac index obstruction","AdS/CFT correspondence"],"falsifier":"Choose a Lovelock coupling vector $\\alpha$ with $A_1(\\alpha)=0$ that still admits an asymptotically hyperbolic solution, and solve the formal Fefferman–Graham recursion to order $n-1$; if the recursion is underdetermined there or forces a power $x^\\beta$ outside the monoid generated by the indicial roots, then the polyhomogeneity theorem as stated fails for that $\\alpha$. The paper's hypothesis $1\\in\\operatorname{LimSec}(\\alpha)$ excludes this case by definition, so the open question is whether the stated theorem can be freed from it.","tokens_in":37468,"feed_emoji":"📐","tokens_out":11591,"duration_ms":112218,"temperature":0.7,"pith_summary":"This paper establishes that conformally compact metrics satisfying the Lovelock equations—a family of higher-curvature generalizations of the Einstein equation—admit polyhomogeneous expansions near their conformal boundary. That means the compactified metric can be written in powers of a defining function $x$ and, where needed, powers of $\\log x$, so the formal Fefferman–Graham expansion is genuinely realized rather than being a formal device. In even boundary dimensions the expansion develops a logarithmic term whose leading coefficient is a trace-free tensor built from the boundary Schouten tensor, generalizing the ambient obstruction tensor of the Einstein case. The same methods produce a formal solution to the singular Yamabe-$(2q)$ problem and a Dirac-index obstruction to filling a spin manifold by such metrics when the scalar curvature is bounded below.","feed_headline":"Lovelock metrics gain real boundary expansions","feed_subtitle":"Higher-curvature Einstein generalizations realize the Fefferman–Graham expansion, with a log-term obstruction in even dimensions.","key_machinery":"The engine is the Lovelock tensor $F_\\alpha(g)=\\sum_q \\alpha_q\\big[(\\operatorname{Ric}_g^{(2q)}-\\lambda_{(2q)}g)-\\frac{\\alpha_q}{2q}(\\operatorname{scal}_g^{(2q)}-(n+1)\\lambda_{(2q)})g\\big]$, a symmetric divergence-free polynomial in the Riemann tensor whose constants $\\lambda_{(2q)}$ make hyperbolic space a solution. To get ellipticity, the paper uses the modified tensor $Q_\\alpha(g,t)=F_\\alpha(g)-\\Phi_\\alpha(g,t)$, where the gauge-fixing term $\\Phi_\\alpha$ cancels the Bianchi terms in the linearization; the resulting principal operator is $A_1(\\alpha)/4[-(n-1)(\\Delta+2n)(C_g(r)g)+2(\\Delta-2)r_0]+O(x^{N+1})$. The coefficient $A_1(\\alpha)=\\sum_q \\alpha_q(-\\tfrac12)^{q-1}\\frac{(n-2)!(2q)!}{2(n-2q)!}$ is what makes the operator elliptic. The argument then uses the known indicial roots of $\\Delta+2n$ and $\\Delta-2$ on trace and trace-free symmetric two-tensors, together with Green's operators $G_\\infty$ and $G_0$ adapted to weighted Hölder and polyhomogeneous spaces, to construct corrections at an increasing sequence of weights $\\mu_++a_k$; when an indicial root is hit, a logarithmic term is inserted exactly as in the Einstein case.","core_discovery":"The central claim, Theorem 6.1, is that any asymptotically hyperbolic Lovelock metric with a smooth conformal infinity is polyhomogeneous in a collar neighborhood of the boundary. Concretely, with $g=(dx^2+h_x)/x^2$, the family $h_x$ has an expansion $h_x=h_0+h_2x^2+\\cdots+h_{n-1}x^{n-1}+h_nx^n+\\cdots$ when $n$ is odd, and $h_x=h_0+h_2x^2+\\cdots+h_{n,1}x^{n-1}\\log x+h_nx^n+\\cdots$ when $n$ is even. The paper proves that every such metric actually has this expansion, not merely that a formal expansion can be written down. The argument fixes a gauge by modifying the Lovelock tensor with a Bianchi-gauge term so that its linearization at an asymptotically hyperbolic metric is a Laplace-type operator with principal part $A_1(\\alpha)[-(n-1)(\\Delta+2n)(C_g(r)g)+2(\\Delta-2)r_0]/4$; invertibility of the two Laplace pieces, control of indicial roots, and Green's operators on weighted spaces then build an approximating sequence whose remainder decays to all orders. The paper also computes the leading-order term of the even-dimensional obstruction tensor and derives formal solutions and filling obstructions from the same expansion machinery.","pith_inferences":["Beyond the paper, the same Green-operator induction should apply to any divergence-free curvature-tensor equation whose gauge-fixed linearization is a sum of Laplace-type operators with computable indicial roots; the Lovelock case is a testbed, not the only possible one.","Beyond the paper, the obstruction-tensor formula implies a family of conformal boundary invariants indexed by $\\alpha$: when $A_1(\\alpha)$ changes sign or vanishes, the locus of smoothly extendable conformal classes can move, so different higher-curvature theories may disagree on which conformal infinities are smoothly fillable.","Beyond the paper, the filling obstruction likely extends to other spinorial index invariants or nonzero eta invariants, and the scalar-curvature bound $\\operatorname{scal}_g\\ge -n(n+1)$ is probably not optimal; this is an extension, not a claim of the paper.","Beyond the paper, the singular Yamabe-$(2q)$ obstruction $L^{(2q)}$, which transforms with conformal weight $-(n+1)$, is a natural candidate for a Lovelock analogue of $Q$-curvature; computing it explicitly in low dimensions would be a concrete test of that analogy."],"forward_implications":["Every asymptotically hyperbolic Lovelock metric with smooth conformal infinity has a genuine polyhomogeneous expansion; the formal Fefferman–Graham expansion computed by earlier deformation arguments is therefore the actual expansion of the metric.","In even boundary dimension, the coefficient of $x^{n-1}\\log x$ is controlled by the obstruction tensor $\\mathcal{O}=\\frac{A_1(\\alpha)}{c_n}\\Delta^{n/2-2}(P^k_{ij,k}-P^k_{k,ij})$, so smoothness of the expansion is governed to leading order by the vanishing of that tensor built from the boundary Schouten tensor.","When $\\tilde B_{1,2}(\\beta,\\kappa)\\neq 0$, the singular Yamabe-$(2q)$ problem has a formal solution $u=x+u_2x^2+\\cdots+u_{n+1}x^{n+1}+L^{(2q)}x^{n+2}\\log x$, making the linear combination of scalar-$(2q)$ curvatures vanish up to $O(x^{n+2}\\log x)$.","For a compact spin manifold of dimension $4k$ with a boundary metric of positive Yamabe invariant, a nonvanishing Dirac index $I(X,h)$ forbids a conformally compact Lovelock filling with $\\operatorname{scal}_g\\geq -n(n+1)$.","Odd-dimensional conformal infinities admit smooth expansions, while even-dimensional ones carry the logarithmic obstruction; this dichotomy matches the Einstein setting."],"supporting_citations":[{"why":"It supplies the formal Fefferman–Graham expansion for Lovelock metrics and the linearization of the Lovelock tensor that the regularity argument upgrades to polyhomogeneity.","marker":"[Alb20]"},{"why":"It provides the gauge-fixing, Green's operator, and double-weighted Hölder space machinery that the paper adapts from the Einstein case.","marker":"[BH14]"},{"why":"It gives existence of Einstein metrics with prescribed conformal infinity on the ball and the indicial root computations for Laplace operators on symmetric two-tensors.","marker":"[GL91]"},{"why":"It supplies the Fredholm theory and indicial root analysis for Laplace operators on asymptotically hyperbolic manifolds used to establish invertibility.","marker":"[Lee06]"},{"why":"It gives the ambient obstruction tensor in the Einstein case whose leading-order term Theorem 7.2 generalizes to Lovelock metrics.","marker":"[GH05]"},{"why":"It establishes the formal singular Yamabe expansion framework that the paper extends to the scalar-$(2q)$ problem.","marker":"[Gra17]"},{"why":"It supplies the $I$-invariant, gluing lemmas, and vanishing lemma that the paper uses to obstruct conformally compact Lovelock fillings.","marker":"[GHS21]"},{"why":"It defines the Fefferman–Graham ambient metric construction whose boundary expansion the polyhomogeneity result realizes.","marker":"[FG12]"}],"fun_headline_variants":["Lovelock metrics prove polyhomogeneous expansions","Even-dimensional log obstruction for Lovelock metrics","Fefferman-Graham expansion realized in Lovelock theory","Lovelock equations yield full boundary asymptotics","Boundary logs block smoothness in even Lovelock metrics"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof needs the nondegeneracy condition $A_1(\\alpha)\\neq 0$ (equivalently $1\\in \\operatorname{LimSec}(\\alpha)$), which makes the two leading-order equations of Section 6.2 linearly independent and keeps the gauge-fixed linearization elliptic; Theorems 1.1 and 6.1 are stated without this hypothesis, but Section 6.2 invokes it to start the induction, so if a Lovelock combination had $A_1=0$ the argument as written would not go through.","fun_headline_variants_meta":{"raw":{"variants":["Lovelock metrics prove polyhomogeneous expansions","Even-dimensional log obstruction for Lovelock metrics","Fefferman-Graham expansion realized in Lovelock theory","Lovelock equations yield full boundary asymptotics","Boundary logs block smoothness in even Lovelock metrics"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000301,"raw_usage":{"total_tokens":1736,"prompt_tokens":943,"completion_tokens":793,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":559,"completion_tokens_details":{"reasoning_tokens":716}},"tokens_in":559,"tokens_out":793,"duration_ms":8455,"temperature":1.0,"reasoning_tokens":716,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:31:03.196192+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Choose a Lovelock coupling vector $\\alpha$ with $A_1(\\alpha)=0$ that still admits an asymptotically hyperbolic solution, and solve the formal Fefferman–Graham recursion to order $n-1$; if the recursion is underdetermined there or forces a power $x^\\beta$ outside the monoid generated by the indicial roots, then the polyhomogeneity theorem as stated fails for that $\\alpha$. The paper's hypothesis $1\\in\\operatorname{LimSec}(\\alpha)$ excludes this case by definition, so the open question is whether the stated theorem can be freed from it.","supporting_citations":[],"review_version":1}