{"id":"3ff6a917-a6ca-40fa-bb97-b661688c033e","arxiv_id":"2501.13822","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"An auto-stabilized weak Galerkin method, replacing stabilizers with bubble functions, is analyzed and tested for elasticity interface problems on nonconvex polytopal meshes.","lead":"This paper introduces a stabilizer-free weak Galerkin finite element scheme for elasticity interface problems on nonconvex polygonal and polyhedral meshes. The method claims optimal-order error estimates and is tested numerically on smooth interface problems.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central stability estimate Lemma 4.7 is imported without proof from the authors' unpublished companion [63]; uniqueness and Theorem 6.3 collapse if its h-independent norm equivalence fails, and the experiments use r1=r2=k+1/k+2, not the analyzed r1=r2=2N+k-1.","rationale":"The paper's objective is to prove optimal-order convergence of a stabilizer-free weak Galerkin scheme for elasticity interface problems on nonconvex polytopal meshes. For that conclusion to hold, the local discrete weak strain and weak divergence must control the boundary jumps of the weak functions; that is precisely the content of Lemma 4.7. Tracing the proof flow confirms this is load-bearing: Theorem 4.9 uses (4.7) to infer ||w||_{1,h}=0 from a(w,w)=0, and Theorem 6.3 uses (4.7) to convert the error-equation bound (6.8) into the |||v|||-control needed for the final estimate. Without the lower bound C1||v||_{1,h} <= |||v|||, neither uniqueness nor the stated convergence rate follows. The lemma is not proved in this manuscript and is attributed to [63], an unpublished companion preprint by the same authors; this is a genuine verifiability gap. The difficulty is not merely stylistic: for nonconvex cells the squared bubble Phi_B vanishes on interior supporting lines of edges, and the edge-bubble inequalities (4.5)-(4.6) require an unspecified extension of vb into T. These issues may be fixable, and the numerical results suggest the method is viable, but they use r1=r2=k+1 or k+2 rather than the analyzed r1=r2=2N+k-1, so they do not validate the theorem as stated. The reader's CONDITIONAL verdict is therefore appropriate; the proposed local generalized-eigenvalue check would settle whether Lemma 4.7 holds in the generality claimed.","tokens_in":20509,"tokens_out":10545,"duration_ms":101432,"concrete_test":"Verify Lemma 4.7 locally on a family of scaled nonconvex polygons. For a fixed concave 5-edge polygon and for an L-shaped element, assemble the element matrices for A_T(v,w)=a_T(v,w) and B_T(v,w) corresponding to the three terms defining ||v||_{1,h} on V(k,T), using the theoretical degrees r1=r2=2N+k-1. Compute the generalized eigenvalues of the pencil A_T versus B_T for a shape-regular refinement sequence h=1,1/2,...,1/64, and check whether the minimal eigenvalue is bounded below by a constant independent of h and the maximal eigenvalue is bounded above. If the ratio deteriorates as h decreases, the h-independent norm equivalence in Lemma 4.7 fails and Theorem 6.3 is unsubstantiated. Repeating the same computation with r1=r2=k+1 or k+2 would also show whether the numerical experiments test a different, unproved parameter regime.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 4.9 (uniqueness) and Theorem 6.3 (the optimal-order error estimate) both rest on Lemma 4.7, the two-sided norm equivalence C1||v||_{1,h} <= |||v||| <= C2||v||_{1,h} for v in V_h. The lemma is stated as from [63] with no proof, and [63] is the same authors' arXiv:2411.17879, so the decisive stability input is not verifiable from this manuscript. The proposed mechanism is also delicate for nonconvex polytopes: the squared edge-distance bubble Phi_B = product of l_i^2 can vanish on codimension-one lines passing through the interior of a concave element, and the edge-bubble estimates (4.5)-(4.6) require a polynomial extension of the boundary data vb into T that is not specified. If the constants C1 and C2 depend on the number of edges N or on h through the nonconvex geometry in a way not covered by the cited shape regularity, then a(w,w)=|||w|||^2 no longer controls ||w||_{1,h}; consequently the conclusion w=0 in Theorem 4.9 and the bound (6.7) in Theorem 6.3 would fail. Independent numerical support is absent: the experiments in Tables 7.1-7.8 set r1=r2=k+1 or k+2, whereas Lemmas 4.2-4.7 are stated for r1=r2=2N+k-1 on nonconvex cells, so the computed examples do not exercise the analyzed regime.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript introduces an auto-stabilized weak Galerkin (WG) finite element method for elasticity interface problems on polytopal meshes without convexity constraints. The method replaces the traditional stabilizer with bubble functions, leading to a symmetric positive-definite formulation. The paper claims optimal-order error estimates in the discrete H1-norm under sufficient smoothness of the exact solution, and reports numerical experiments on nonconvex polygonal grids that show optimal convergence rates.","tokens_in":20872,"tokens_out":3967,"duration_ms":34000,"significance":"If the central stability and approximation lemmas were established, the method would be a meaningful contribution: it removes the stabilizer from WG methods for elasticity while allowing nonconvex polytopal meshes, which is of practical and theoretical interest. The manuscript contains a clear derivation of the error equation and a standard error-estimate argument. Its main weaknesses are that the decisive lemmas are imported without proof from the authors' own unpublished preprint [63], and the numerical experiments use lower weak-derivative degrees than the analysis requires, so the computations do not validate the analyzed scheme. The self-referential chain of preprints [41,42,63] also makes the paper not self-contained.","major_comments":[{"comment":"These lemmas are stated as coming from [63] with proofs omitted. Since [63] is an arXiv preprint by the same authors (arXiv:2411.17879) and not a published source, the uniqueness proof (Theorem 4.9) and the main error estimate (Theorem 6.3) rest on unverifiable imported results. In particular, Lemma 4.7 (the norm equivalence) is load-bearing: if it fails, the method is not stable and both Theorem 4.9 and Theorem 6.3 collapse. The manuscript should provide complete proofs for these lemmas in an appendix or cite a peer-reviewed source.","section":"§4, Lemmas 4.2–4.7 and §5, Lemma 5.1"},{"comment":"The numerical experiments do not test the analyzed method. The text states that the computations use r1 = r2 = k+1 on the grids of Figure 7.1 and r1 = r2 = k+2 on the 7-edge nonconvex polygons of Figure 7.3, whereas Definition 4.1 and Lemma 4.7 for nonconvex elements require r1 = r2 = 2N + k − 1. The observed optimal rates are therefore not evidence for Theorem 6.3 in the nonconvex setting. The authors should rerun the experiments with the analyzed parameter choice, or extend the analysis to cover the lower-degree regime actually used.","section":"§7.2–§7.6, Tables 7.1–7.8"},{"comment":"For a nonconvex polytope, the bubble function Φ_B = ∏ l_i^2 can vanish on codimension-one subsets passing through the interior (for example, a line through a reentrant vertex and another edge), so the assertion that Φ_B ≥ ρ0 on a subdomain T̂ is not obvious and may be false. The edge-based bubble φ_{e_i} has the same issue. Since Lemmas 4.5–4.7 rely on these bubble functions, the geometric claim must be proved or an additional assumption on the polytopes must be imposed.","section":"Definition 4.1 and Lemmas 4.5–4.7"},{"comment":"The estimates (4.5) and (4.6) compare a boundary quantity involving vb with an L2 norm over T, but the extension of vb from ∂T to T is not defined. The proof of Lemma 6.2 likewise uses boundary traces of Q0u − Qbu without specifying the polynomial extension into the element. This gap is part of the imported lemmas, but because it is essential to the error analysis, it should be addressed explicitly in the revision.","section":"Lemmas 4.5–4.6 and Lemma 6.2"}],"minor_comments":[{"comment":"The phrase 'where where' appears in Examples 7.2, 7.3, 7.4, 7.5, and 7.6; it should be corrected.","section":"§7.2–§7.6"},{"comment":"The word 'ploytopal' should be 'polytopal'.","section":"Key words"},{"comment":"The heading 'Auto-Stablized' should be 'Auto-Stabilized'.","section":"Algorithm 3.1"},{"comment":"The sentence describing r1 and r2 for Figure 7.1 contains a duplication ('r1 = r2 = k + 1 and r1 = r2 = k + 1'); it should be clarified which value applies to triangles and which to nonconvex polygons.","section":"§7.2, first sentence"},{"comment":"The trace inequality for polynomials is quoted with an unspecified constant; since r1 and r2 depend on the number of edges N, the constant may depend on N, and this dependence is not discussed.","section":"§4, inequality (4.2)"}],"recommendation":"major_revision","confidential_remarks":"The paper's central stability and approximation lemmas are drawn from the authors' own unpublished arXiv preprints [41,42,63], and the numerical tests do not exercise the parameter regime analyzed. These are fixable in a revision if the authors supply the missing proofs and rerun the experiments with the prescribed r1 = r2 = 2N + k − 1. The geometric degeneracy of bubble functions on nonconvex polytopes also deserves serious attention. I recommend major revision rather than rejection, because the overall framework is plausible and the missing pieces are identifiable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things. The genuinely new piece is the interface error equation and the optimal-order estimate for a stabilizer-free WG scheme on nonconvex polytopal meshes. The stability of the method, however, rests on Lemma 4.7, whose proof is not here—it is quoted from the same authors' unpublished companion [63]—and the numerical experiments never use the polynomial degrees the theory requires.\n\nWhat the paper does well: the scheme is clean and symmetric positive definite; the error equation in Lemma 5.2 is derived carefully and the interface-jump terms on the right-hand side are handled properly; Theorem 6.3 follows a standard interpolation-plus-duality structure once the norm equivalence is granted. The experiments cover several nonconvex grid families and large coefficient contrasts, and the observed rates match the expected order. Credit is earned there.\n\nThe soft spots are load-bearing. Lemma 4.7 is a two-sided norm equivalence between the discrete H1 seminorm and the energy norm for the bubble-enriched weak space. Uniqueness and the error estimate both collapse if its constants are not h-independent. It is stated as coming from [63] with no proof. That is not automatically fatal—the companion may be correct—but it makes the central claim unverifiable from this manuscript. Also, the theory sets r1=r2=2N+k-1 for nonconvex cells, while the experiments use r1=r2=k+1 or k+2. So the tables test a cheaper variant, not the analyzed method, and the discrepancy is not discussed. The bubble construction for nonconvex elements deserves scrutiny too: the squared-distance bubble can vanish on codimension-one lines inside a concave cell, and Lemmas 4.5–4.6 require a polynomial extension of boundary data that is not specified. Those details need to be nailed down in the companion.\n\nThis paper is for people working on stabilizer-free WG or polytopal FEM for interface problems. If the companion holds, it is a solid incremental contribution. As it stands, the paper is not self-contained and the experiments validate the wrong degree regime. I would not desk-reject it. Send it to referees with a request that the authors either include the proofs of Lemmas 4.2–4.7 or make the companion's proofs clearly accessible, and that they rerun the experiments with the analyzed degree combination or explain why the lower degrees are adequate. My vote is conditional acceptance after major revision.","headline":"Stabilizer-free WG for elasticity interfaces has a new error analysis, but the decisive stability lemma comes from an unpublished companion and the numerics test a different degree regime.","tokens_in":21408,"tokens_out":2926,"would_cite":false,"duration_ms":24376,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65N30","65N15","65N12","74N20","35B45","35J50","35J35"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper builds an auto-stabilized weak Galerkin method for elasticity interface problems whose bubble-enriched weak operators remove the need for stabilizers and yield optimal-order convergence on nonconvex polytopal meshes.","keywords":["weak Galerkin","auto-stabilized","bubble functions","elasticity interface problems","nonconvex polytopal meshes","weak strain tensor","optimal-order error estimates","finite element methods"],"falsifier":"On a nonconvex polygonal element like the 5-edge cells of the paper's type-2 grids, set $v_0=0$ and choose a nonzero boundary polynomial $v_b$, compute the ratio $\\|v\\|_{1,h}/\\vert\\vert\\vert v\\vert\\vert\\vert$ on a sequence of meshes refined toward that element, and check whether the ratio stays bounded as $h\\to0$; unbounded growth would falsify Lemma 4.7 and with it the stability proof.","tokens_in":20232,"feed_emoji":"🧮","tokens_out":14283,"duration_ms":103279,"temperature":0.7,"pith_summary":"The paper introduces a weak Galerkin finite element scheme for elasticity interface problems that needs no stabilizer term, the usual extra penalty in WG methods, and works on polygonal or polyhedral meshes whose elements may be nonconvex. Auto-stabilization comes from computing the discrete weak strain tensor and discrete weak divergence with polynomials of degree $r_1=r_2=2N+k-1$, where $N$ is the number of faces of the element, so bubble functions provide the stability that stabilizers previously supplied. The central result is that the scheme has a unique solution and satisfies the optimal-order error estimate $\\vert\\vert\\vert u-u_h\\vert\\vert\\vert \\le C h^k (\\sum_i \\|u\\|^2_{k+1,\\Omega_i})^{1/2}$ in the discrete energy norm. The payoff is a simpler, symmetric, positive definite scheme that extends WG to nonconvex polytopal meshes, and the numerical experiments on nonconvex polygonal grids confirm the predicted convergence orders.","feed_headline":"Bubble functions replace stabilizers in WG elasticity solver","feed_subtitle":"Bubble-enriched weak operators keep the scheme symmetric, positive definite, and optimal-order accurate on nonconvex meshes.","key_machinery":"The central objects are the element-based bubble function $\\Phi_B=\\prod_{i=1}^N l_i^2(x)$, which vanishes on the boundary of a polytopal element and stays bounded below on an interior subdomain, and the edge/face-based bubble functions $\\varphi_{e_i}$, which vanish on all but one face and control the boundary mismatch $v_b-v_0$. These functions replace stabilizers through the discrete weak strain tensor $\\epsilon_{w,r_1,T}$ and discrete weak divergence $\\nabla_{w,r_2,T}\\cdot$, computed with $r_1=r_2=2N+k-1$ on nonconvex elements. The load-bearing mechanism is Lemma 4.7's norm equivalence, $C_1\\|v\\|_{1,h}\\le\\vert\\vert\\vert v\\vert\\vert\\vert\\le C_2\\|v\\|_{1,h}$, which converts the energy norm into a norm that sees both the strain of the interior value and the boundary discrepancy, enabling uniqueness and the optimal error bound.","core_discovery":"The paper's claim is that Algorithm 3.1 — the space $V_h$ of interior polynomials $[P_k(T)]^d$ together with boundary values from $[P_{k-1}(e)]^d$ plus rigid motions, paired with the bilinear form $a(u,v)=\\sum_T (2\\mu\\epsilon_w u,\\epsilon_w v)_T+(\\lambda\\nabla_w\\cdot u,\\nabla_w\\cdot v)_T$ and the interface jump data — defines a symmetric, positive definite linear system with a unique solution. The error analysis shows that, whenever the exact solution lies in $\\prod_{i=1}^N [H^{k+1}(\\Omega_i)]^d$, the discrete energy error obeys $\\vert\\vert\\vert u-u_h\\vert\\vert\\vert \\le C h^k (\\sum_i \\|u\\|^2_{k+1,\\Omega_i})^{1/2}$. The proof runs through the error equation of Lemma 5.2, the approximation estimates of Lemma 6.2, and the norm equivalence Lemma 4.7, which the paper takes from reference [63]; the theory is corroborated by $P_1$, $P_2$, and $P_3$ computations on two families of nonconvex polygonal meshes.","pith_inferences":["The paper does not explore whether the same bubble-enriched weak operators eliminate stabilizers for other interface problems, such as Stokes flow, Maxwell equations, or biharmonic equations; the self-stabilization mechanism described here suggests a testable extension in those settings.","The superconvergent rates observed on grids that mix triangles with nonconvex polygons indicate that on such special meshes the actual error may be better than $O(h^k)$; identifying which mesh features produce the extra order would require a refined analysis beyond this paper.","Because the bubble degree $r_1=r_2=2N+k-1$ grows with face count, the practical trade-off between the simpler formulation and the higher-degree weak-operator computation on high-face-count polyhedra remains to be quantified; a cost comparison with stabilizer-based WG would settle it."],"forward_implications":["On any shape-regular polytopal mesh, including nonconvex elements, the $P_k$ auto-stabilized scheme converges in the discrete energy norm at the optimal rate $O(h^k)$ whenever the exact solution is piecewise $H^{k+1}$.","Implementations of weak Galerkin elasticity solvers can drop the stabilizer terms entirely, reducing code complexity while keeping a symmetric positive definite system.","Because the operators, bubble functions, and norm equivalence are formulated for $d=2,3$, the same construction carries over to polyhedral meshes in three dimensions.","The global number of unknowns and the sparsity pattern are unchanged from stabilizer-based WG; only the polynomial degrees $r_1=r_2$ used in computing weak operators are raised.","The theory requires $u\\in\\prod_{i=1}^N [H^{k+1}(\\Omega_i)]^d$, so optimal order is guaranteed for smooth material regions and interfaces, not for interface singularities below that regularity."],"supporting_citations":[{"why":"Supplies the proofs of the bubble-function lemmas, above all Lemma 4.7's norm equivalence, on which uniqueness and the error estimate rest.","marker":"[63]"},{"why":"Introduces the weak strain tensor and the WG elasticity formulation that Algorithm 3.1 adapts into its stabilizer-free form.","marker":"[60]"},{"why":"Provides the locking-free weak Galerkin elasticity framework and the discrete weak strain and weak divergence operators used in the bilinear form.","marker":"[52]"},{"why":"Supplies the shape-regularity condition, trace inequalities, and polynomial approximation estimates used throughout the error analysis.","marker":"[57]"},{"why":"Supports Lemma 4.2's bubble-function control of the interior strain by the weak strain on polytopal meshes without convexity constraints.","marker":"[41]"},{"why":"Supports the bubble-function tools for higher-order polytopal elements used in the stability estimates of Section 4.","marker":"[42]"}],"fun_headline_variants":["Bubble functions replace stabilizers in WG elasticity","Auto-stabilized WG: bubbles make stabilizers obsolete","No stabilizers needed: bubbles stabilize WG for nonconvex meshes","WG elasticity without stabilizers, using bubbles"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's existence, uniqueness, and error estimates all rest on Lemma 4.7, imported from reference [63] without proof: on nonconvex polytopal elements the discrete energy norm and the companion norm that sees interior strain plus boundary mismatch must remain equivalent with constants independent of the mesh size.","fun_headline_variants_meta":{"raw":{"variants":["Bubble functions replace stabilizers in WG elasticity","Auto-stabilized WG: bubbles make stabilizers obsolete","No stabilizers needed: bubbles stabilize WG for nonconvex meshes","WG elasticity without stabilizers, using bubbles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001407,"raw_usage":{"total_tokens":5672,"prompt_tokens":918,"completion_tokens":4754,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":534,"completion_tokens_details":{"reasoning_tokens":4689}},"tokens_in":534,"tokens_out":4754,"duration_ms":26222,"temperature":1.0,"reasoning_tokens":4689,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T15:35:25.266582+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a nonconvex polygonal element like the 5-edge cells of the paper's type-2 grids, set $v_0=0$ and choose a nonzero boundary polynomial $v_b$, compute the ratio $\\|v\\|_{1,h}/\\vert\\vert\\vert v\\vert\\vert\\vert$ on a sequence of meshes refined toward that element, and check whether the ratio stays bounded as $h\\to0$; unbounded growth would falsify Lemma 4.7 and with it the stability proof.","supporting_citations":[{"cited_title":"Simplified Weak Galerkin Methods for Linear Elasticity on Nonconvex Domains","cited_arxiv_id":"2411.17879","evidence_quote":"Supplies the proofs of the bubble-function lemmas, above all Lemma 4.7's norm equivalence, on which uniqueness and the error estimate rest."},{"cited_title":"W ang and S","cited_arxiv_id":null,"evidence_quote":"Introduces the weak strain tensor and the WG elasticity formulation that Algorithm 3.1 adapts into its stabilizer-free form."},{"cited_title":"Auto-Stabilized Weak Galerkin Finite Element Methods on Polytopal Meshes without Convexity Constraints","cited_arxiv_id":"2408.11927","evidence_quote":"Supports Lemma 4.2's bubble-function control of the interior strain by the weak strain on polytopal meshes without convexity constraints."},{"cited_title":"Auto-Stabilized Weak Galerkin Finite Element Methods for Biharmonic Equations on Polytopal Meshes without Convexity Assumptions","cited_arxiv_id":"2409.05887","evidence_quote":"Supports the bubble-function tools for higher-order polytopal elements used in the stability estimates of Section 4."}],"review_version":1}