{"id":"af359b0c-31c8-46b8-bcb1-1f091fa23bfe","arxiv_id":"2412.20188","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Weak solutions of an anisotropic nonlocal cross-diffusion system converge to weak solutions of the corresponding local cross-diffusion system in the vanishing viscosity limit.","lead":"This paper proves that a nonlocal anisotropic cross-diffusion system with a viscosity parameter converges, as the parameter vanishes, to a local cross-diffusion system. The result matters because anisotropic diffusion is used in tissue growth and population dynamics models, and this fills a gap in the mathematical theory of such limits.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"In §6 the weak form of Eq. (3) is written with flux ∫∇ϕ·A∇mν instead of ∫ nν∇ϕ·A∇mν; consequently the limit equation (11) is not the weak form of ∂n0/∂t=∇·(n0A∇n0)+…, so the proof of Theorem 1.1 is incomplete as written.","rationale":"I read the paper in good faith. The intended result—that the nonlocal anisotropic system localizes to (4)—is plausible and the entropy-dissipation strategy is coherent, and the paper contains a genuinely nontrivial proof of the entropy identity and the compactness. However, the text as provided contains a concrete mathematical inconsistency at the point where the limit equation for the total population is derived. This is more immediately load-bearing than the existence claim: Theorem 1.1 is a conditional statement, so an unproven existence theorem makes it vacuous but not false, whereas the weak-form error makes the proof of the convergence invalid even under the existence hypothesis. I therefore cannot sign off on the central claim as written. I note that a corrected weak form would likely restore the argument, so a CONDITIONAL verdict remains appropriate. I disagree with the reader's choice of weakest assumption—the Section 6 inconsistency is, to me, the single most load-bearing concern—but I do not move the verdict.","tokens_in":15968,"tokens_out":21984,"duration_ms":192181,"concrete_test":"Sum the weak forms in Definition 2.1 for i=1,2 and compare with the displayed weak form at the start of Section 6; verify whether the flux term contains the factor nν. If it does, check that passing to the limit yields ∫ n0∇ϕ·A∇n0, not ∫∇ϕ·A∇n0, in Eq. (11); then confirm the entropy identity (12) by re-deriving it from ∂n0/∂t = ∇·(n0A∇n0)+… rather than from the printed (11).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The displayed weak form at the start of Section 6 (above Eq. (11)) reads −∫ nν∂tϕ −∫φ(0)nin +∫∇ϕ·A∇mν = ∫... . Summing the two equations in Definition 2.1 for i=1,2 gives the total-density weak form with flux +∫ nν∇ϕ·A∇mν, not +∫∇ϕ·A∇mν. Passing to the limit in the erroneous form yields Eq. (11) with +∫∇ϕ·A∇n0, which is the weak form of ∂n0/∂t = ∇·(A∇n0)+…, whereas the text states the limit equation as ∂n0/∂t = ∇·(n0A∇n0)+…. This is an internal inconsistency in the proof of the central claim: the derivation of the limit total-density equation, and hence the entropy identity (12) on which the strong convergence of ∇mν in §7 relies, is not justified. The error is independent of the existence gap identified by the reader: even assuming the claimed existence of weak solutions, the convergence proof fails at this step unless the missing factor nν is inserted. The fix is straightforward, but as written the main theorem is not established.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the singular limit ν→0 in a nonlocal anisotropic cross-diffusion system (2): two continuity equations for species densities n_ν^(1), n_ν^(2), with a common velocity field −A∇m_ν, coupled through the anisotropic Brinkman equation −ν∇·(A∇m_ν)+m_ν=n_ν. The main result, Theorem 1.1, asserts that along a subsequence the nonlocal solutions converge to a weak solution of the local system (4), with strong L2 convergence of the total density n_ν and of the velocity gradient ∇m_ν. The proof strategy is to obtain ν-uniform estimates, prove an entropy identity for weak solutions of the nonlocal system, exploit the entropy dissipation to get compactness of m_ν and n_ν, pass to the limit in the total-density equation, and finally compare the entropy identities for ν>0 and ν=0 to upgrade weak convergence of ∇m_ν to strong convergence, which allows passage to the limit in the equations for the individual species.","tokens_in":16321,"tokens_out":23048,"duration_ms":247811,"significance":"If the proof is carried out correctly, the result fills a genuine gap: localisation limits for anisotropic degenerate cross-diffusion systems are not covered by the existing isotropic literature. The mollified entropy equality in Section 4 and the anisotropic velocity regularity in Section 2.1 are useful building blocks, and the argument is genuinely analytical with no fitted parameters or reverse-engineered quantities. However, the manuscript in its current form contains several load-bearing gaps, most notably an incorrect weak form of the limit total-density equation and an unproved existence/L∞-bound assertion. These issues are local and repairable, so the paper is not beyond revision, but it cannot be accepted as written.","major_comments":[{"comment":"The weak form displayed at the beginning of Section 6 is not the weak form of Eq. (3): the flux term is written as ∫∇φ·A∇m_ν dx dt, whereas summing the two equations in Definition 2.1 gives ∫ n_ν∇φ·A∇m_ν dx dt. Consequently the limiting equation (11) contains ∫∇φ·A∇n_0 instead of ∫ n_0∇φ·A∇n_0, i.e. it is the weak form of a linear diffusion equation and not of ∂_t n_0 = ∇·(n_0A∇n_0)+… as stated immediately afterwards. This error propagates into the entropy identity (12), because that identity is derived from the wrong limit equation. The proof of Theorem 1.1 is therefore incomplete at this step. The correction is local: restore the missing factors n_ν and n_0 in the two flux terms, and then pass to the limit using Lemma 5.2 together with ∇m_ν ⇀ ∇n_0 in L2.","section":"Section 6, Eq. (11)"},{"comment":"Existence of weak solutions to System (2) for fixed ν is asserted but not proved, and the asserted regularity A∇m ∈ L2(0,T;H1) inherits the same status. More importantly, Lemma 3.1 supplies the ν-uniform L∞ bound 0≤n_ν≤n̄, but its proof is explicitly formal: it selects a maximum point (x*,t*) of n_ν and uses pointwise relations such as ∂_t n_ν=0 and ∇n_ν=0, which are not available for the weak solutions in Definition 2.1 (they are only known to have n ∈ C([0,T];L2)). Because the uniform L∞ bound underpins Lemmas 3.2–3.6, Corollary 4.2, and hence the whole compactness argument, this is a load-bearing gap. The authors should either give a complete approximation argument establishing existence and the L∞ bound, or state the existence theorem with precise hypotheses that explicitly include the bound.","section":"Section 2 and Lemma 3.1"},{"comment":"The displayed chain of equalities in Lemma 5.2 is not correct as written. From Brinkman’s equation one has m_ν−n_ν=ν∇·(A∇m_ν), so ||m_ν−n_ν||^2_{L2} = ν^2||∇·(A∇m_ν)||^2_{L2} = −ν∫ n_ν∇·(A∇m_ν) dx dt − ν∫ ∇m_ν·A∇m_ν dx dt. The second displayed line in the proof appears to have the wrong sign and to omit the factor ν on the second term. The bound ≤ Cν also needs an argument: it follows from the entropy identity (9) and the relation −∫ n_ν∇·(A∇m_ν) = ∫ ∇m_ν·A∇m_ν + ν||∇·(A∇m_ν)||^2, together with boundedness of −∫ n_ν∇·(A∇m_ν). This chain should be written out. Since Lemma 5.2 is the only place proving n_ν→n_0 strongly in L2, the gap is central.","section":"Lemma 5.2"}],"minor_comments":[{"comment":"The keyword list ('Inviscid Limit, Brinkman-to-Darcy Limit, Tissue Growth') does not match the nonlocal approximation content of the paper and should be updated to reflect the anisotropic nonlocal-to-local limit studied here.","section":"Abstract and keywords"},{"comment":"The notation n_ν^(i) ⇀* n_0^(i) in L∞(0,T;L1∩L∞) is not standard, since L1∩L∞ is not a dual space in an obvious way; the authors should state the convergence more precisely as weak-* in L∞(0,T;L∞) and weak in L1.","section":"Theorem 1.1"},{"comment":"There are typographical inconsistencies in the test functions: Eq. (11) writes φ(0) where Definition 2.1 uses ϕ(x,0), and the text switches between ϕ and φ without comment.","section":"Section 6, Eq. (11)"},{"comment":"The convergence React_ν→React_0 is delegated to references [16, Section 4.2] and [22, Section 4]; since at that stage only the total density converges strongly while the individual densities converge weakly, a short justification in the anisotropic setting would improve readability and completeness.","section":"Section 7"}],"recommendation":"major_revision","confidential_remarks":"The paper has a solid core idea and the main theorem is likely recoverable after the corrections outlined above. The Section 6 weak-form error and the Lemma 5.2 algebra are the most serious; the existence and L∞-bound gap in Section 2/Lemma 3.1 should be addressed head-on, preferably by a complete approximation argument or a precise statement of the imported theorem. I do not see a reason to reject, but the current version is not acceptable in its present form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper extends the known isotropic localization limit for this cross-diffusion system to an anisotropic Darcy-type coupling, and that is a genuine gap in the literature. The entropy-dissipation identity for arbitrary weak solutions (Prop 4.1) and the comparison of entropy identities to upgrade weak convergence of ∇mν to strong are the right tools, and the regularity estimates for the anisotropic velocity field in Section 2.1 are new. The overall strategy is coherent and, I think, likely correct.\n\nBut the manuscript as written has two gaps that need attention before it can be accepted. First, existence of weak solutions to the anisotropic nonlocal system is asserted without proof; the paper simply says a standard approximation scheme from [1,16,22,25] adapts. That may be true, but it is not demonstrated, and the L∞ bound in Lemma 3.1 is a formal maximum principle applied to weak solutions, with no justification. Second, and more seriously, the displayed weak form in Section 6 is wrong: it reads ∫∇ϕ·A∇mν, but the total density flux is nνA∇mν, so the correct term is ∫ nν∇ϕ·A∇mν. Consequently the limit equation (11) is not the weak form of ∂n0/∂t=∇·(n0A∇n0); it would correspond to ∇·(A∇n0). The error is internal: the entropy identity (12) and the comparison in Section 7 rely on the intended equation, so the proof as written is incomplete. The fix is straightforward—insert the missing nν and use strong convergence of nν—so this is a repairable flaw, not a fatal one, but it has to be corrected.\n\nThe citations look appropriate; the anisotropic case is explicitly left open in the cited works. Self-citation is heavy but justified here. The paper is for readers working on cross-diffusion systems and nonlocal-to-local limits, and it deserves a serious referee. I would send it to review, with the request that the authors fix the weak-form error and either prove existence or clearly state it as an assumption.\n\nBest","headline":"A genuinely new anisotropic localization limit with a sound entropy-comparison strategy, but the manuscript as written has a missing factor in the Section 6 weak form that breaks the central proof until fixed.","tokens_in":16734,"tokens_out":2567,"would_cite":false,"duration_ms":24826,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35K57","47N60","35B45","35K55","35K65","35Q92"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the anisotropic nonlocal cross-diffusion system converges, as the viscosity parameter goes to zero, to a local degenerate cross-diffusion system, using an entropy-dissipation identity that holds for every weak…","keywords":["Inviscid Limit","Brinkman-to-Darcy Limit","Tissue Growth","Anisotropic Parabolic-Hyperbolic Cross-Diffusion Systems","entropy dissipation identity","nonlocal-to-local limit","weak solutions","degenerate parabolic systems"],"falsifier":"Take the simplest nontrivial case, $G^{(i)}\\equiv 0$ and constant elliptic tensor $A$, and solve the nonlocal system (2) numerically on a bounded interval with smooth initial data; the entropy identity (9) should hold up to discretization error, and $\\|\\nabla m_\\nu-\\nabla n_0\\|_{L^2}$ should tend to zero as $\\nu\\to0$. A persistent violation of the entropy identity, or a sequence of weak solutions for which $\\int_0^T\\int n_\\nu|\\nabla m_\\nu|^2$ is not bounded uniformly in $\\nu$, would falsify the paper's central convergence claim.","tokens_in":15775,"feed_emoji":"🧮","tokens_out":9170,"duration_ms":85820,"temperature":0.7,"pith_summary":"This paper proves a vanishing-viscosity (localisation) limit for a two-species anisotropic cross-diffusion system from population dynamics. The system couples the species densities $n^{(i)}_\\nu$ to a velocity potential $m_\\nu$ through an anisotropic Brinkman equation, and the theorem shows that as $\\nu \\to 0$ a subsequence of weak solutions converges to a weak solution of the corresponding local degenerate system. The proof rests on an entropy-dissipation identity that is shown to hold for every weak solution of the nonlocal system, giving uniform bounds that lead to strong convergence of the total density and of the anisotropic velocity field. If the theorem is correct, the local anisotropic model with Darcy-type relation $v = A\\nabla p$ is recovered as the limit of a regularised nonlocal model, a case left open by earlier works.","feed_headline":"Nonlocal-to-local limit proven for anisotropic diffusion","feed_subtitle":"A new entropy-dissipation identity, valid for every weak solution, carries the proof and covers the anisotropic Darcy case.","key_machinery":"The key object is the Boltzmann-Shannon entropy $H[n]=\\int n(\\log n-1)\\,dx$ and the entropy-dissipation identity (Proposition 4.1), which holds for any weak solution: $H[n_\\nu(T)]-H[n_{\\rm in}] - \\int_0^T\\int n_\\nu\\nabla\\cdot(A\\nabla m_\\nu)\\,dx\\,dt = \\int_0^T\\int \\log n_\\nu(n^{(1)}_\\nu G^{(1)}(n_\\nu)+n^{(2)}_\\nu G^{(2)}(n_\\nu))\\,dx\\,dt$. The identity is obtained by mollifying the total-density equation and controlling the DiPerna-Lions commutator; together with the anisotropic Brinkman equation it yields the uniform dissipation bound that controls $\\sqrt{n_\\nu}\\nabla m_\\nu$ and $\\nabla m_\\nu$ in $L^2$. The same machinery is then applied to the limiting equation, and comparing the two entropy identities gives the upper semi-continuity of the quadratic dissipation $\\int\\nabla m_\\nu\\cdot A\\nabla m_\\nu$, which combined with weak lower semi-continuity forces strong $L^2$ convergence of $\\nabla m_\\nu$.","core_discovery":"The central discovery is that the anisotropic version of the nonlocal-to-local limit is valid at the level of weak solutions. For each $\\nu>0$ the paper considers weak solutions of $\\partial_t n^{(i)}_\\nu = \\nabla\\cdot(n^{(i)}_\\nu A\\nabla m_\\nu)+n^{(i)}_\\nu G^{(i)}(n_\\nu)$ coupled with $-\\nu\\nabla\\cdot(A\\nabla m_\\nu)+m_\\nu=n_\\nu$, where $A$ is symmetric and uniformly elliptic and $G^{(i)}$ are decreasing homeostatic growth functions. Theorem 1.1 states that, up to a subsequence, $n^{(i)}_\\nu$ converges weak-$*$ in $L^\\infty(0,T;L^1\\cap L^\\infty)$, the total density $n_\\nu=n^{(1)}_\\nu+n^{(2)}_\\nu$ converges strongly in $L^2$, $m_\\nu$ converges weakly in $L^2(0,T;H^1)$ and strongly in $L^2$, and the limit pair $(n^{(1)}_0,n^{(2)}_0)$ is a weak solution of the local system $\\partial_t n^{(i)}_0=\\nabla\\cdot(n^{(i)}_0 A\\nabla n_0)+n^{(i)}_0 G^{(i)}(n_0)$ with $n_0=n^{(1)}_0+n^{(2)}_0$. The proof upgrades weak convergence of $\\nabla m_\\nu$ to strong convergence by comparing the entropy identity for the nonlocal system with the entropy identity for the limit, using the fact that $A$ defines an inner product.","pith_inferences":["The same two-entropy comparison argument should extend to more than two species or to pressure laws with different nonlinearities, as long as a total-density entropy identity and an elliptic coupling equation are available.","Because the existence of weak solutions is only asserted and cited, a fully self-contained version of Theorem 1.1 would need a proof that the parabolic approximation schemes produce solutions satisfying the uniform $L^\\infty$ and regularity bounds; the convergence argument itself is independent of which scheme is used.","A quantitative rate for the convergence $m_\\nu\\to n_0$ and $\\nabla m_\\nu\\to\\nabla n_0$ might be obtained by tracking the $\\nu$-dependence in the entropy comparison, but the paper does not pursue this.","Numerically, the strong convergence of $\\nabla m_\\nu$ suggests that computing the limiting local model through the nonlocal regularisation with small $\\nu$ should give accurate gradients, not just accurate densities."],"forward_implications":["The anisotropic Darcy-type case $v=A\\nabla p$, which earlier existence results excluded because of strong assumptions on the diffusion matrix, is now covered: the nonlocal system converges to the local degenerate cross-diffusion system (4).","The entropy-dissipation identity is available for any weak solution, so the proof does not need extra regularity beyond the natural weak formulation and can be reused in other anisotropic parabolic-hyperbolic systems.","The strong convergence of $\\nabla m_\\nu$ and of $n_\\nu$ means the velocity field of the limiting population model is genuinely obtained from the nonlocal approximations, not merely in a distributional sense.","The nonlocal Brinkman-regularised system can serve as a well-posed approximating model for the local anisotropic system, since any sequence of its weak solutions has a subsequence converging to a weak solution of the local system."],"supporting_citations":[{"why":"Supplies the isotropic predecessor of the inviscid-limit result and the entropy-comparison strategy adapted here to the anisotropic setting.","marker":"[16]"},{"why":"Provides the approximation procedure and entropy-convergence details cited when the paper asserts existence of weak solutions to the nonlocal system.","marker":"[22]"},{"why":"Handles the Brinkman-to-Darcy inviscid limit with nonlinear pressure and contributes the entropy-identity approach used in the proof.","marker":"[25]"},{"why":"Supplies the DiPerna-Lions mollification and commutator framework used to justify the entropy equality for weak solutions.","marker":"[21]"},{"why":"Provides the specific commutator lemma (Lemma 2.3) used to pass the mollified entropy calculation to the limit.","marker":"[38]"},{"why":"Cited for compactness and approximation results underpinning the asserted existence of weak solutions to the nonlocal system.","marker":"[1]"}],"fun_headline_variants":["Anisotropic nonlocal-to-local limit proven via entropy identity","Entropy identity proves anisotropic nonlocal-to-local limit","Nonlocal-to-local limit for anisotropic cross-diffusion","Weak solution entropy identity yields anisotropic limit","Entropy dissipation identity unlocks anisotropic limit"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that for every $\\nu>0$ the nonlocal system has at least one weak solution with the uniform bounds $0\\le n_\\nu\\le\\bar n$, $n_\\nu\\in C([0,T];L^2)$, and $A\\nabla m_\\nu\\in L^2(0,T;H^1)$; the existence statement is asserted in Section 2 with references to approximation schemes, not proved in this paper, and the $L^\\infty$-bound in Lemma 3.1 is justified by a formal maximum-principle argument that may not apply directly to weak solutions.","fun_headline_variants_meta":{"raw":{"variants":["Anisotropic nonlocal-to-local limit proven via entropy identity","Entropy identity proves anisotropic nonlocal-to-local limit","Nonlocal-to-local limit for anisotropic cross-diffusion","Weak solution entropy identity yields anisotropic limit","Entropy dissipation identity unlocks anisotropic limit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000995,"raw_usage":{"total_tokens":4220,"prompt_tokens":953,"completion_tokens":3267,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":569,"completion_tokens_details":{"reasoning_tokens":3194}},"tokens_in":569,"tokens_out":3267,"duration_ms":24225,"temperature":1.0,"reasoning_tokens":3194,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T23:28:13.411845+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the simplest nontrivial case, $G^{(i)}\\equiv 0$ and constant elliptic tensor $A$, and solve the nonlocal system (2) numerically on a bounded interval with smooth initial data; the entropy identity (9) should hold up to discretization error, and $\\|\\nabla m_\\nu-\\nabla n_0\\|_{L^2}$ should tend to zero as $\\nu\\to0$. A persistent violation of the entropy identity, or a sequence of weak solutions for which $\\int_0^T\\int n_\\nu|\\nabla m_\\nu|^2$ is not bounded uniformly in $\\nu$, would falsify the paper's central convergence claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the DiPerna-Lions mollification and commutator framework used to justify the entropy equality for weak solutions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the specific commutator lemma (Lemma 2.3) used to pass the mollified entropy calculation to the limit."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Cited for compactness and approximation results underpinning the asserted existence of weak solutions to the nonlocal system."},{"cited_title":"David, T","cited_arxiv_id":null,"evidence_quote":"Supplies the isotropic predecessor of the inviscid-limit result and the entropy-comparison strategy adapted here to the anisotropic setting."},{"cited_title":"From Finite to Continuous Phenotypes in (Visco-)Elastic Tissue Growth Models","cited_arxiv_id":"2409.02904","evidence_quote":"Provides the approximation procedure and entropy-convergence details cited when the paper asserts existence of weak solutions to the nonlocal system."},{"cited_title":"On the inviscid limit connecting Brinkman's and Darcy's models of tissue growth with nonlinear pressure","cited_arxiv_id":"2306.03752","evidence_quote":"Handles the Brinkman-to-Darcy inviscid limit with nonlinear pressure and contributes the entropy-identity approach used in the proof."}],"review_version":1}