{"id":"bd802189-0e05-4030-9f8e-217edc9023a8","arxiv_id":"2411.19760","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Local insensitizing controllability is established for a quasilinear volume-surface reaction-diffusion system with dynamic boundary conditions.","lead":"This paper proves that a quasilinear reaction-diffusion equation on a volume and its boundary can be made insensitive to small unknown initial perturbations by a localized control, provided the data are small. It is the first insensitizing controllability result for quasilinear equations with dynamic (Wentzell-type) boundary conditions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Prop 5.3 uses boundary-in-time conditions in P opposite to those required by the Carleman estimate (5.11), so the key observability inequality (5.19) is not established as written.","rationale":"The reader accepted the paper and identified the imported Carleman estimates as the weakest assumption. My review finds a more concrete, internal issue in the proof of Proposition 5.3: the time boundary conditions defining the space P are opposite to those used in the Carleman estimate and in the variational derivation. This is not a criticism of the cited external estimates; it is a mismatch inside the submitted argument. The issue is load-bearing because Proposition 5.3 is the only bridge from the linearized cascade system to the surjectivity of Λ'(0,0,0), which is required before applying Lyusternik--Graves. The likely fix is to reverse the boundary conditions in P, and the rest of the proof appears consistent with that correction. I therefore recommend CONDITIONAL rather than REJECT: the manuscript should be revised to correct the definition of P and verify that (5.19) and the Euler--Lagrange identities hold under the corrected conditions. If the authors confirm the correction, the central claim is probably sound; until then, the proof as written has a genuine gap.","tokens_in":44148,"tokens_out":23670,"duration_ms":212181,"concrete_test":"Redo Proposition 5.3 with P replaced by P' = {(Y,Z): y(·,T)=0, z(·,0)=0}. First check that the observability inequality (5.19) follows by applying Proposition 5.2 to (Y,Z) as the solution of (5.2) with sources f1 = L*_1Y - Z1_O, g1 = L1Z, f1Γ = L*_2Y - ZΓ1_Σ, g1Γ = L2Z. Second, re-derive from the Riesz identity (5.21) that Ψ = μ0^{-2}(L*_1Φ - K1_O, L*_2Φ - KΓ1_Σ) satisfies L1Ψ = F + v1_ω and H satisfies the backward equation with Ψ(0)=0 and H(T)=0. If both steps close, the text's boundary conditions are a typographical reversal and the theorem is likely correct; if either fails, Proposition 5.3 and hence the surjectivity of Λ'(0) are not proved.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The core of the linear controllability argument is Proposition 5.3, which proves surjectivity of Λ'(0) via a Fursikov--Imanuvilov variational construction. The space P is defined (Section 5.1, before (5.21)) by y(x,0)=0 and z(x,T)=0. However, Proposition 5.2 and its Carleman estimate (5.11) are proved for solutions of the adjoint system (5.2), whose time boundary conditions are φ(·,T)=0 and k(·,0)=0. The proof of Proposition 5.2 explicitly uses these conditions when constructing the cut-off functions on (T/2,T), with (y(·,T),yΓ(·,T))=(0,0) and (z(·,0),zΓ(·,0))=(0,0). For (Y,Z) in the space P as written, these are reversed, so the estimate (5.19) \"∫ μ^{-2}|y|² + ∫ μ^{-2}|z|² ≤ C B((Y,Z),(Y,Z))\" does not follow from (5.11). Independently, integrating B((Φ,K),(Y,Z)) by parts with y(0)=0 and z(T)=0 leaves boundary terms at t=T, which would force a terminal condition on the forward state rather than the required initial condition Ψ(0)=0. Since the surjectivity claim for Λ'(0) rests entirely on Proposition 5.3, this is a load-bearing gap in the proof of the main theorem.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a local insensitizing controllability result for a quasilinear volume-surface reaction-diffusion equation with dynamic boundary conditions of generalized Wentzell type. The main theorem (Theorem 3.1) asserts that for sufficiently small data F satisfying the weighted smallness condition (3.1), there exists a control v in H^1(0,T;L^2(ω)) ∩ L^2(0,T;H^2(ω)) that insensitizes the functional J in (1.2). The proof reformulates insensitizing controllability as null controllability of a cascade quasilinear system (Lemma 3.2), proves null controllability of the linearized cascade via a Carleman estimate and a Fursikov–Imanuvilov variational argument (Propositions 5.2 and 5.3), derives weighted estimates for the linearized states (Proposition 5.4), and then applies the Lyusternik–Graves inverse mapping theorem to the nonlinear mapping Λ. The paper is long and detailed, with several auxiliary results proved in appendices, but the central proof currently contains a time-boundary condition mismatch in the variational argument that is load-bearing for the surjectivity of Λ'(0).","tokens_in":44405,"tokens_out":10022,"duration_ms":87195,"significance":"If the main theorem is correct, this is the first insensitizing controllability result for a quasilinear parabolic equation with dynamic boundary conditions, and it generalizes earlier semilinear results such as [11] to a volume-surface setting in which the functional includes both bulk and surface energy. The paper is ambitious and the strategy is natural: reduce insensitizing to null controllability of a cascade, prove Carleman-based linear controllability, and use a local inversion theorem. The authors provide explicit weighted estimates and state regularity of the controls, and they include proofs of the nonlinear well-posedness and of the reduction lemma in appendices. However, the proof of the key linear surjectivity step is not currently sound as written, so the significance claim is contingent on a substantive repair. The paper deserves serious consideration after the gap is fixed.","major_comments":[{"comment":"The space P is defined with y(x,0)=0 and z(x,T)=0, but the Carleman estimate (5.11) in Proposition 5.2 is proved for solutions of the adjoint system (5.2), whose time boundary conditions are φ(·,T)=0 and k(·,0)=0. In the proof of Proposition 5.2, these boundary conditions are used explicitly when constructing Y=ϑΦ and Z=ϑK on (T/2,T), with (y(·,T),yΓ(·,T))=(0,0) and (z(·,0),zΓ(·,0))=(0,0). Therefore the coercivity estimate (5.19), ∫ μ^{-2}(|y|²+|z|²) plus surface terms ≤ C B((Y,Z),(Y,Z)), does not follow from (5.11) for the elements of P as written. Since (5.19) is the only argument that B is a scalar product and that the linear form F is continuous, the surjectivity of Λ′(0) claimed in §6.2 is not established. This is a load-bearing gap. It can likely be repaired by changing the conditions on P to y(x,T)=0 and z(x,0)=0, which are the natural conditions for the operators L*_1 and L_1 appearing in B, but the proof must be rewritten accordingly and all subsequent steps rechecked with this correction.","section":"§5.1, Proposition 5.3 and Proposition 5.2"},{"comment":"The step 'from (5.21), (Ψ,H) is the unique distributional solution of (3.5) associated with the control v' is asserted without a displayed verification. After defining Ψ, H and v by (5.23), one must show, by taking variations in (5.21) with smooth compactly supported test functions, that the Euler–Lagrange equation is exactly the four equations of the linearized cascade (3.5) with the control v=−χμ1^{-2}φ|ω. This is a standard duality computation, but it is load-bearing for the claim that the constructed control actually drives the cascade to zero. The paper should present this computation in detail, especially since the preceding boundary-condition issue makes it essential to verify which time boundary conditions the test functions satisfy.","section":"§5.1, equations (5.21)–(5.23)"}],"minor_comments":[{"comment":"Several estimates in Proposition 5.4 are only sketched; for example, (5.42) is obtained by 'the same multiplication techniques as in (5.41)', and the passage from (5.49) to (5.50) suppresses intermediate terms involving sup μ4²||Ψt||². These are standard energy estimates, but the section would be easier to verify if the corresponding H-estimates were displayed or the omitted terms were identified explicitly.","section":"§5.2, Proposition 5.4"},{"comment":"The statement of Proposition 5.2 says 'for any Φ=(φ,φΓ), K=(k,kΓ) ∈ ET' but does not state the time boundary conditions φ(·,T)=0 and k(·,0)=0 that are used in the proof. The statement should include these conditions, otherwise the estimate (5.11) is not true as written.","section":"§5.1, Proposition 5.2"},{"comment":"There is a typo in the proof of Proposition 4.5: 'Υ′(0) is suejctive' should read 'surjective'.","section":"Appendix A"}],"recommendation":"major_revision","confidential_remarks":"The main theorem relies on Carleman estimates imported from [23] and [24], both coauthored by one of the present authors. This is not by itself a problem, but the boundary-condition mismatch found in §5.1 suggests that the applicability of those estimates to the variational space P needs independent scrutiny. The variational identification after (5.23) should also be made fully explicit. If the authors can repair the time-boundary conditions in Proposition 5.3 and verify the Euler–Lagrange step, the paper would be a solid contribution; in its present form, the central surjectivity claim is not established."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nBottom line: this is a genuine, first-of-its-kind result — insensitizing controls for a quasilinear volume-surface reaction-diffusion equation with dynamic boundary conditions, with the cost covering both bulk and surface energy — and it is built with the standard toolkit (Carleman estimate for the linearized cascade, then Lyusternik-Graves), executed in detail over 47 pages. But the reader's ACCEPT is too generous as written: Proposition 5.3, the load-bearing linear controllability step, contains a boundary-condition error.\n\nThe space P in Prop 5.3 is defined with y(x,0)=0 and z(x,T)=0. The Carleman estimate (5.11) it invokes is proved for solutions of the adjoint system (5.2), which have φ(·,T)=0 and k(·,0)=0. In the bilinear form B, Y plays the φ role (backward operator L*) and Z plays the k role (forward operator L), so applying (5.11) to (Y,Z) requires y(T)=0 and z(0)=0 — the reverse of P as written. So (5.19) is not established. The same swap breaks the variational identification: integrating B by parts with P's conditions leaves boundary terms that force Ψ(T)=0 and H(0)=0 instead of the required Ψ(0)=0 and H(T)=0. The weight μ₀⁻² vanishing at t=0 does save Ψ(0)=0, but the forced terminal condition on the forward state is overdetermined, and the terminal condition on H is lost. Since the surjectivity of Λ'(0) rests entirely on Prop 5.3, Theorem 3.1 is not proven as written.\n\nThe fix looks cosmetic — swap the conditions to y(T)=0 and z(0)=0 and the pieces line up — which suggests a slipped sign or typo rather than a deep flaw, though the corrected argument needs a careful re-read. The paper does a number of things well that point the same way: the literature placement is honest, the insensitizing-to-cascade reduction (Lemma 3.2) is proved cleanly in an appendix, and the quasilinear well-posedness (Prop 4.5) is actually addressed rather than assumed. Two smaller notes: several estimates in Prop 5.4 are waved through with 'using similar arguments' after long computations, and the Carleman estimates are imported from the authors' own earlier work [23],[24] — not a flaw, but the core estimate gets no independent reproof here.\n\nWho this is for: researchers in insensitizing controls or Carleman-based null controllability for parabolic systems with dynamic boundary conditions. Recommendation: send to peer review; a serious referee should verify a corrected Prop 5.3 in full. The result is probably true, but acceptance should wait for the correction.","headline":"Genuine first quasilinear insensitizing result for volume-surface systems with dynamic boundary conditions, but the Fursikov–Imanuvilov space in Prop 5.3 has time boundary conditions reversed relative to the Carleman estimate it uses, so the surjectivity proof breaks as written.","tokens_in":45000,"tokens_out":34835,"would_cite":true,"duration_ms":253830,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35K55","35K57","35K59","93C20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims local insensitizing controllability for quasilinear volume-surface reaction-diffusion equations with dynamic boundary conditions, proved by reducing the problem to null controllability of a cascade system and applying a…","keywords":["insensitizing controls","quasilinear parabolic equations","dynamic boundary conditions","volume-surface reaction-diffusion","null controllability","Carleman estimates","Lyusternik-Graves theorem","cascade system"],"falsifier":"Compute the weighted observability inequality (5.3) in the one-dimensional case, where the tangential operators vanish: if the decoupled bulk-heat-with-trace system admits data for which the inequality fails with the stated weights, then Proposition 5.3 and Theorem 3.1 would fail. A direct numerical test for a small T and a shrinking observation set ω ∩ O would also help settle whether the technical assumption (A3) is actually necessary.","tokens_in":43882,"feed_emoji":"🎯","tokens_out":5568,"duration_ms":47483,"temperature":0.7,"pith_summary":"The paper claims that a quasilinear volume-surface reaction-diffusion equation with dynamic boundary conditions admits local insensitizing controls: controls that render an energy functional insensitive to small unknown perturbations of the initial state. The proof reduces insensitization to null controllability of a cascade quasilinear system, proves null controllability for its linearization via Carleman estimates with weights blowing up at t=0, and then applies a local inversion theorem. If Theorem 3.1 is correct, this is the first insensitizing controllability result for quasilinear systems covering both the bulk and the surface with dynamic boundary conditions. The control found has regularity $H^{1}$(0,T;$L^{2}$(ω)) ∩ $L^{2}$(0,T;$H^{2}$(ω)) and its norm is bounded by weighted norms of the source term.","feed_headline":"Insensitizing controls reach quasilinear surface-diffusion equations","feed_subtitle":"Small-data controls now work for volume-surface reaction-diffusion with dynamic boundary conditions.","key_machinery":"The argument is carried by a Carleman estimate for the adjoint of the linearized cascade system with dynamic boundary conditions (Lemma 5.1 and Proposition 5.2); here a Carleman estimate is a weighted a priori inequality, with weights blowing up as t→0^+, that bounds volume and surface integrals of the adjoint state by localized bulk observation terms. This estimate is imported from known Carleman results for dynamic boundary conditions, then adapted to the cascade and used to prove null controllability and weighted estimates on the control and state (Propositions 5.3 and 5.4). The nonlinear step is the mapping Λ defined in (3.4) between weighted Hilbert spaces X and Y; proving that Λ is $C^{1}$ and that Λ′(0,0,0) is surjective lets the Lyusternik-Graves inverse mapping theorem deliver the local solution.","core_discovery":"The central discovery is that insensitizing controllability for the quasilinear volume-surface equation is equivalent to null controllability at time zero of a cascade system, where an adjoint variable must start at (0,0). The authors construct a nonlinear mapping Λ whose surjectivity at the origin is exactly the null controllability of the linearized cascade system, and they prove that surjectivity using a Carleman estimate for the coupled volume-surface adjoint system together with weighted energy estimates. The Lyusternik-Graves inverse mapping theorem then yields a local solution of Λ(Ψ,H,v)=(F,0) for sufficiently small weighted data F, forcing the adjoint variable to vanish at t=0 and thereby insensitizing the original functional.","pith_inferences":["The same linearization-plus-local-inversion scheme should extend to diffusion and reaction coefficients depending on the state and its gradient, as the authors themselves note; the main technical work would be verifying that the nonlinear estimates in Section 6 still close under weaker coefficient regularity.","If a global Carleman estimate could be proved without the overlap condition ω ∩ O ≠ ∅, the rest of the argument appears independent of that geometry, so the main theorem would likely carry over to disjoint control and observation regions.","The weighted smallness condition (3.1) demands exponential decay of the source as t→0^+; a testable direction would be to weaken the weights and see which data still lie in the range of Λ′(0), potentially broadening the class of admissible sources."],"forward_implications":["If Theorem 3.1 holds, every sufficiently small source term satisfying the weighted bound (3.1) admits a control that insensitizes the functional, with the control enjoying H^1(0,T;L^2(ω)) ∩ L^2(0,T;H^2(ω)) regularity.","The result is the first insensitizing controllability statement for quasilinear systems with dynamic boundary conditions, covering nonlinearities both in the bulk and on the surface.","For the relaxed functional J_{t_0} starting at t_0>0, small initial data and source terms can be insensitized with an L^2(ω_T) control by first steering the state to zero at t_0 and then applying the theorem from t_0 onward.","The geometric overlap condition ω ∩ O ≠ ∅ is used essentially in the Carleman proof, so the disjoint control-observation case is left open by this argument.","The proof works only in dimensions d ≤ 3, leaving the result open for d ≥ 4."],"supporting_citations":[{"why":"Supplies the Carleman estimate for parabolic equations with dynamic boundary conditions used in Lemma 5.1 for d ≥ 2.","marker":"[23]"},{"why":"Supplies the one-dimensional Carleman estimate used in Lemma 5.1 for d = 1.","marker":"[24]"},{"why":"Provides the weighted-estimate method, including weights blowing up at t = 0, used to prove null controllability of the linearized cascade system.","marker":"[25]"},{"why":"Provides the Lyusternik-Graves inverse mapping theorem that the paper applies to the nonlinear mapping Λ.","marker":"[19]"},{"why":"Earlier null controllability result for the volume-surface equation with dynamic boundary conditions, used for local solutions and for the relaxed-functional argument.","marker":"[21]"},{"why":"First insensitizing control result for semilinear parabolic equations with dynamic boundary conditions, serving as the baseline the present result extends to the quasilinear setting.","marker":"[11]"},{"why":"Insensitizing controls for a class of quasilinear parabolic equations with standard boundary conditions, showing the gap the paper fills for dynamic boundary conditions.","marker":"[8]"}],"fun_headline_variants":["Quasilinear surface-diffusion insensitivity via cascade null control","Insensitizing dynamic boundary equations via cascade controllability","Volume-surface reaction-diffusion: insensitivity from null control","Small-data insensitizing controls for surface-diffusion systems","Cascade proof establishes insensitivity in quasilinear diffusion"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole proof rests on the global Carleman estimate that is imported from two cited papers for the coupled volume-surface linearized adjoint system; if that estimate fails or does not apply to this coupling, the surjectivity of the linearized map and hence the main theorem collapses.","fun_headline_variants_meta":{"raw":{"variants":["Quasilinear surface-diffusion insensitivity via cascade null control","Insensitizing dynamic boundary equations via cascade controllability","Volume-surface reaction-diffusion: insensitivity from null control","Small-data insensitizing controls for surface-diffusion systems","Cascade proof establishes insensitivity in quasilinear diffusion"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00062,"raw_usage":{"total_tokens":2785,"prompt_tokens":763,"completion_tokens":2022,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":379,"completion_tokens_details":{"reasoning_tokens":1939}},"tokens_in":379,"tokens_out":2022,"duration_ms":13984,"temperature":1.0,"reasoning_tokens":1939,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T05:51:04.813761+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the weighted observability inequality (5.3) in the one-dimensional case, where the tangential operators vanish: if the decoupled bulk-heat-with-trace system admits data for which the inequality fails with the stated weights, then Proposition 5.3 and Theorem 3.1 would fail. A direct numerical test for a small T and a shrinking observation set ω ∩ O would also help settle whether the technical assumption (A3) is actually necessary.","supporting_citations":[{"cited_title":"Maniar, M","cited_arxiv_id":null,"evidence_quote":"Supplies the Carleman estimate for parabolic equations with dynamic boundary conditions used in Lemma 5.1 for d ≥ 2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the one-dimensional Carleman estimate used in Lemma 5.1 for d = 1."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the weighted-estimate method, including weights blowing up at t = 0, used to prove null controllability of the linearized cascade system."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Lyusternik-Graves inverse mapping theorem that the paper applies to the nonlinear mapping Λ."},{"cited_title":"Et-tahri, S.-E","cited_arxiv_id":null,"evidence_quote":"Earlier null controllability result for the volume-surface equation with dynamic boundary conditions, used for local solutions and for the relaxed-functional argument."},{"cited_title":"Zhang, J","cited_arxiv_id":null,"evidence_quote":"First insensitizing control result for semilinear parabolic equations with dynamic boundary conditions, serving as the baseline the present result extends to the quasilinear setting."},{"cited_title":"Liu, Insensitizing controls for a class of quasilinear parabolic eq uations, Journal of Differential Equations , vol","cited_arxiv_id":null,"evidence_quote":"Insensitizing controls for a class of quasilinear parabolic equations with standard boundary conditions, showing the gap the paper fills for dynamic boundary conditions."}],"review_version":1}