{"id":"5faff0a9-d454-4985-a9a4-48d826e89b1a","arxiv_id":"2412.19428","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Uniform measure attractors exist uniquely for non-autonomous McKean-Vlasov stochastic reaction-diffusion equations on unbounded thin domains and converge to the attractor of the collapsed limit equation as the thickness tends to zero.","lead":"This paper proves that a class of McKean-Vlasov stochastic reaction-diffusion equations on unbounded thin domains has unique uniform measure attractors, and that these attractors converge upper-semicontinuously to the attractor of the limit equation on R^n as the domain collapses. A generalist might read it to see how long-time statistical behavior of distribution-dependent stochastic PDEs behaves under dimensional reduction.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.1 invokes Theorem 2.2 whose joint-continuity hypothesis is not established: Lemma 5.1 only proves continuity on bounded subsets of P4(L2(O))×H(g0), and the proof's Vitali step needs uniform 4th-moment bounds that can fail for weakly convergent measures.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern: the mismatch between the full joint-continuity hypothesis of Theorem 2.2 and the bounded-set continuity actually proved in Lemma 5.1. This concern is genuine and is the main obstacle to Theorem 5.1. I do not share the reader's secondary worry about 'false weak-closure' of the moment ball: the functional μ↦∫||x||^4 dμ is lower semicontinuous for weak convergence, so the limit ν in Lemma 5.3 does inherit the fourth-moment bound from the absorbing set. The tail-estimate and tightness arguments in Section 4 and Lemma 5.3 appear plausible, and Theorem 6.1 is conditional on the existence of Aε; if the existence gap is closed, the upper-semicontinuity proof is largely in place. Because the gap may be repairable by proving full continuity, by citing a weakened criterion, or by a direct attractor argument using only continuity on the absorbing set, the conditional verdict is appropriate. No change to the reader's verdict is needed, but the paper should address the bounded-to-full continuity gap explicitly before the central theorem can be considered established.","tokens_in":37475,"tokens_out":20364,"duration_ms":186227,"concrete_test":"Analytical check: take a scalar analogue of (3.21) with drift containing the μ(||·||^2) term from (A1), and choose initial laws μn=(1−1/n)δ0+(1/n)δn, so μn→δ0 weakly but μn(||·||^2)=n→∞. Using the Gronwall estimate (5.3)–(5.8), derive a lower bound on d_P(U^{g,ε}(t,0)μn, U^{g,ε}(t,0)δ0) for fixed small t>0. If this distance does not converge to 0, then U^{g,ε}(t,τ) is not jointly continuous on all of P4(L2(O))×H(g0), confirming that Theorem 2.2 cannot be invoked as written. If the distance does converge to 0 for this example, test with μn=(1−1/n)δ0+(1/n)δ_{a_n} for a_n→∞ slower, and check whether a non-vanishing limit appears, which would still demonstrate the proof gap.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central proof step is Theorem 5.1, which states that by Lemma 5.1 the process is 'jointly continuous over bounded of P4(L2(O)) and H(g0)' and then applies Theorem 2.2. However, Theorem 2.2 (together with Definition 2.1) requires joint continuity on Pp(X)×H(g0) with no boundedness restriction. Lemma 5.1 proves only a restricted statement: if Lξn→Lξ weakly and E||ξn||^4, E||ξ||^4 ≤ R, and gn→g in H(g0), then L(u^{gn,ε}(t,τ,ξn))→L(u^{g,ε}(t,τ,ξ)). The proof uses Vitali's theorem to upgrade Skorokhod a.s. convergence to L2 convergence, which requires uniform integrability of ||ξn||^2 — exactly the uniform 4th-moment bound. For a sequence μn→μ weakly with μ∈P4(L2(O)), the moments of μn need not be bounded; e.g., μn=(1−1/n)δ0+(1/n)δn → δ0 weakly while μn(||·||^2)=n→∞. For such data the estimate (5.8) gives E||v_n(t)||^2 ≤ E||ξn−ξ||^2 e^{c(t−τ)} + o(1), and the initial L2 error may diverge. Since the coefficients depend on the law through W2 and through the term ψ1(x*)μ(||·||^2) in (A1), the one-step map can genuinely fail to be continuous at such μ. Thus Theorem 5.1 is not a direct consequence of Theorem 2.2 as stated; the paper neither proves full continuity nor supplies a modified criterion (e.g., continuity on the absorbing set) that would make Theorem 2.2 applicable.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies non-autonomous McKean-Vlasov stochastic reaction-diffusion equations on unbounded thin domains O_epsilon that collapse to R^n as epsilon goes to 0. The authors reformulate the equation on a fixed domain, establish uniform a priori estimates including fourth-moment bounds and uniform tail estimates, and introduce a family of processes on the space P_4(L^2(O)) of probability measures with finite fourth moment. The main results are Theorem 5.1, asserting existence and uniqueness of a uniform measure attractor A_epsilon for each epsilon, and Theorem 6.1, asserting upper semicontinuity of these attractors toward the collapsed attractor A_0. The argument relies on a general criterion (Theorem 2.2) imported from a submitted companion paper by the same authors, on joint continuity of the process proved only on bounded subsets, and on asymptotic compactness obtained from uniform tail estimates.","tokens_in":37838,"tokens_out":11354,"duration_ms":104045,"significance":"If the gaps identified below are repaired, the paper would be a useful contribution to the theory of measure attractors for distribution-dependent stochastic PDEs on unbounded domains. Its main positive content is a substantial set of uniform-in-epsilon and uniform-in-symbol estimates, including the fourth-moment and tail estimates that overcome non-compactness of the usual Sobolev embedding on unbounded thin domains. The paper also correctly identifies why the standard dual-semigroup/Feller argument for continuity of the law-evolution process fails for McKean-Vlasov equations. The upper-semicontinuity strategy is natural and plausible, provided the existence theorem is made rigorous. At present, however, the central existence theorem is not proved from the stated hypotheses, so the significance is conditional.","major_comments":[{"comment":"Theorem 5.1 applies Theorem 2.2, whose hypothesis is joint continuity on P_p(X) x H(g0) with no boundedness restriction (Definition 2.1 and Theorem 2.2). Lemma 5.1 proves continuity only on bounded subsets of P_4(L^2(O)) x H(g0): it requires the uniform fourth-moment bounds E||xi^n||^4 <= R and E||xi||^4 <= R. The proof of Theorem 5.1 does not bridge this gap, and the introduction explicitly states that continuity is proved only on the restricted domain (B_{P_4}(r), d_P) x H(g0). To make Theorem 5.1 follow, the authors must either prove full joint continuity on P_4(L^2(O)) x H(g0) or state and prove a modified criterion in which continuity on bounded sets, together with a uniform absorbing set and asymptotic compactness, is sufficient. This issue is load-bearing because Theorem 2.2 is the mechanism that produces the attractor A_epsilon.","section":"Section 5, Theorem 5.1 and Lemma 5.1"},{"comment":"Theorem 2.2, the abstract existence-uniqueness criterion for uniform measure attractors, is quoted without proof from reference [28], a submitted paper by the same three authors. The present manuscript's Theorems 5.1 and 5.2 rest entirely on this external criterion. Since the reader cannot verify the hypotheses or the proof of the criterion, the manuscript is not self-contained at its central point. Please provide a complete proof of Theorem 2.2 in this paper or replace the reference with a published or otherwise publicly verifiable source, and state precisely which topology on P_p(X) is used.","section":"Section 2, Theorem 2.2"},{"comment":"The proof of Theorem 6.1 asserts, with 'By Lemma 4.5', that every mu in A_epsilon satisfies the uniform H^1_epsilon-bound (6.16). Lemma 4.5 gives such a bound only for time-evolved measures U^{g,epsilon}(t,tau)mu with t-tau large; it does not directly transfer to weak limits in A_epsilon. One must argue through complete solutions chi with chi(s)=U^{g,epsilon}(s,tau_n)chi(tau_n) and pass to the limit, using the uniform fourth-moment bound to justify the passage. This argument is absent. Without (6.16), the application of Corollary 6.1 to the full attractor A_epsilon is not justified, so the upper-semicontinuity proof has a gap.","section":"Section 6, Theorem 6.1, Eq. (6.16)"}],"minor_comments":[{"comment":"A concern that the absorbing set K = B_{P_4}(r) is not closed in the weak topology does not land: for mu_n in K with mu_n -> mu weakly, Fatou's lemma applied to the nonnegative continuous function ||x||^4 gives E_mu||x||^4 <= liminf E_{mu_n}||x||^4 <= r^4, so K is closed in (P(L^2(O)), d_P). The step nu in K in Lemma 5.3 is therefore sound.","section":"Lemma 5.3, proof around (5.24)-(5.25)"},{"comment":"The absorbing radius G_1 = M_7/rho_1 appears inconsistent with the norm used in P_4(L^2(O)). Lemma 4.7 gives E||u_epsilon||^4_{H_rho} <= M_7, and since ||u||^2_{L^2} <= (1/rho_1)||u||^2_{H_rho}, the corresponding L^2 fourth-moment bound is M_7/rho_1^2. If rho_1 < 1, the set K defined in (5.9) with G_1 = M_7/rho_1 may not contain the time-evolved measures guaranteed by Lemma 4.7; please correct the constant or clarify the norm convention.","section":"Section 5, Lemma 5.2 and Eq. (5.10)"},{"comment":"The sentence 'which in conjunction with Lemmas 4.1 and 4.1, completes the proof' should refer to Lemmas 4.1 and 4.2.","section":"Section 4, proof of Lemma 4.5"},{"comment":"In the line 'Given nu_epsilon in A_epsilon, since A_epsilon in K', the notation should be A_epsilon subset K rather than A_epsilon in K. Also, Theorem 6.1 states assumptions (A1)-(A3) and (6.1)-(6.2) but omits (3.28), although Lemma 4.5 and Corollary 6.1 rely on it; please add (3.28) to the hypotheses of Theorem 6.1.","section":"Section 6, proof of Theorem 6.1"}],"recommendation":"major_revision","confidential_remarks":"The main technical estimates appear substantial and potentially correct, but the central existence theorem is currently not derivable from the stated results because of the mismatch between the bounded continuity proved in Lemma 5.1 and the full continuity required by the imported Theorem 2.2. The reliance on the unpublished companion paper [28] for the key criterion is also worth weighing in the editorial decision. These issues are fixable, but they require more than local editing."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read. This is the first paper to build uniform measure attractors for McKean-Vlasov reaction-diffusion equations on unbounded thin domains, and to prove upper semicontinuity as ε→0; that is a genuine, publishable-in-principle result in stochastic dynamics on thin domains. The uniform tail estimates and fourth-moment bounds are substantial and mostly carefully derived. The paper also earns credit for admitting in the introduction that its continuity proof is restricted to bounded subsets of P4(L2(O))×H(g0) rather than the whole space.\n\nWhere it breaks down: Theorem 5.1 as stated does not follow. Theorem 2.2 (and Definition 2.1) require joint continuity on Pp(X)×H(g0). Lemma 5.1 only proves continuity under a uniform 4th-moment bound on the initial laws. The proof then says 'bounded continuity plus Lemmas 5.1–5.3 and Theorem 2.2 yields the attractor,' but no argument bridges bounded to full continuity. This is load-bearing. The stress-test example (μ_n=(1−1/n)δ_0+(1/n)δ_n) shows why: the one-step map can genuinely fail to be continuous at a measure whose 4th moment is finite but approximating measures have huge moments. One can likely repair the paper by replacing Theorem 2.2 with a version whose continuity hypothesis is 'joint continuity on the absorbing set' or by proving full weak continuity, but that has to be written.\n\nTwo smaller things. The closedness of K=BP4(L2(O))(G1^{1/4}) in the weak topology is asserted and used in Lemma 5.3; that is false as stated. The conclusion ν∈P4 still follows from liminf/Fatou since all approximating laws have bounded 4th moments, so this is fixable. And the main existence result is imported from the authors' own submitted paper [28]; that is acceptable when the theorem is stated fully, but referees need access to [28] to verify the criterion's hypotheses, and the mismatch above is exactly where that matters.\n\nBottom line: the paper deserves a serious referee. Not because it is flawless—it is not—but because the result is new, the technical machinery is substantial, and the gaps are addressable. My recommendation: send to peer review and require a major revision that either proves the missing continuity or restates and proves a modified criterion.","headline":"First thin-domain uniform measure attractor theorem, but the main theorem's continuity hypothesis is not met; referee it and ask for a repair.","tokens_in":38408,"tokens_out":3083,"would_cite":false,"duration_ms":30106,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B41","37L30","37L55","35B40"],"pacs":[],"model":"deepseek-v4-flash","headline":"For each small thickness, a McKean-Vlasov stochastic reaction-diffusion equation on an unbounded thin domain has a unique uniform measure attractor, and these attractors converge upper-semicontinuously to the limiting equation on R^n as…","keywords":["McKean-Vlasov equation","uniform measure attractor","complete solution","unbounded thin domain","stochastic reaction-diffusion equation","upper semicontinuity","uniform tail estimates","almost periodic external term"],"falsifier":"A concrete observation that would settle the central claim is a sequence of initial laws $\\mu_n$ in $\\mathcal{P}_4(L^2(O))$ with uniformly bounded fourth moments and forcings $g_n\\to g$ in $H(g_0)$ whose solution laws at times $t_n\\to\\infty$ admit no weakly convergent subsequence; that would falsify uniform asymptotic compactness and with it Theorem 5.1. Alternatively, a sequence $\\varepsilon_k\\to0$ and $\\mu_k\\in A_{\\varepsilon_k}$ whose laws do not converge along any subsequence to an element of $A_0\\circ I^{-1}$ would falsify the upper-semicontinuity claim of Theorem 6.1.","tokens_in":37228,"feed_emoji":"📐","tokens_out":9107,"duration_ms":77645,"temperature":0.7,"pith_summary":"This paper establishes that a family of non-autonomous McKean-Vlasov stochastic reaction-diffusion equations on an unbounded thin domain — a strip-like region $O_\\varepsilon$ in $\\mathbb{R}^{n+1}$ that flattens to $\\mathbb{R}^n$ as thickness $\\varepsilon\\to0$ — has a unique uniform measure attractor for each small $\\varepsilon$, and that these attractors vary upper semicontinuously under the collapse. The equations are distribution-dependent: the nonlinearity and noise depend on the law of the solution, so the usual duality between transition semigroups and their adjoints breaks down, and the paper instead proves joint continuity of the law-evolution process on bounded subsets of $\\mathcal{P}_4(L^2(O))$ using regularity and the Vitali theorem. Why this matters: the result identifies the long-time statistical behavior of a genuinely distribution-dependent stochastic PDE on a non-compact, geometrically degenerating domain, and shows it is stable as the geometry degenerates. The proofs rely on uniform tail estimates to overcome the failure of compact Sobolev embedding on unbounded domains.","feed_headline":"Thin-domain attractors survive collapse to R^n","feed_subtitle":"Probability-law attractors for stochastic reaction-diffusion equations exist and converge as the thin domain flattens.","key_machinery":"The argument is carried by a chain of objects. A change of variables $T_\\varepsilon$ transplants the thin domain $O_\\varepsilon$ into the fixed cylinder $O=\\mathbb{R}^n\\times(0,1)$, turning the Laplacian into the weighted operator $A_\\varepsilon$ and the limit equation on $\\mathbb{R}^n$ into $A_0=-\\rho^{-1}\\sum_i(\\rho u_{y_i})_{y_i}$; the average operator $M\\phi(y_*)=\\int_0^1\\phi(y_*,y_{n+1})dy_{n+1}$ and the injection $I$ compare the two geometries. The uniform tail estimates (Lemma 4.6) control the mass of solutions outside large balls uniformly in $\\varepsilon$ and in the forcing $g$, which restores the compactness that the unbounded domain removes from Sobolev embedding. The existence criterion is an imported theorem (Theorem 2.2) asserting that joint continuity of the law processes, a uniform absorbing set, and uniform asymptotic compactness yield a unique uniform measure attractor; the paper verifies these inputs for $\\{U^{g,\\varepsilon}\\}$ in $\\mathcal{P}_4(L^2(O))$.","core_discovery":"On its own terms, the paper's central result is Theorem 5.1 and Theorem 6.1. For each $0<\\varepsilon<\\varepsilon_0$, the family of processes $\\{U^{g,\\varepsilon}(t,\\tau)\\}_{g\\in H(g_0)}$ generated by the transformed equation (3.21) on the fixed domain $O$ possesses a unique uniform measure attractor $A_\\varepsilon$ in $\\mathcal{P}_4(L^2(O))$, equal to the union over the hull $H(g_0)$ of the kernel sections $K^{g,\\varepsilon}(0)$. As $\\varepsilon\\to 0$, these attractors are upper semicontinuous at the collapsed domain: $\\lim_{\\varepsilon\\to0} d_{\\mathcal{P}(L^2(O))}(A_\\varepsilon, A_0\\circ I^{-1})=0$, where $A_0$ is the uniform measure attractor of the limit equation (3.22) on $\\mathbb{R}^n$ and $I$ embeds $\\mathbb{R}^n$ functions as functions constant in the vertical coordinate. In words: every probability law that is selected by the long-time dynamics on the thin domain approaches, as the domain flattens, a law selected by the long-time dynamics of the limiting equation.","pith_inferences":["Beyond the paper, if the bounded-set continuity proved in Lemma 5.1 can be upgraded to full joint continuity, the same proof scheme would apply to other distribution-dependent parabolic SPDEs with fourth-moment bounds, such as distribution-dependent fractional reaction-diffusion equations on unbounded domains.","Beyond the paper, the upper-semicontinuity statement does not by itself give convergence of individual attractor elements; a natural companion question is whether $A_\\varepsilon$ also lower-semicontinuously approximates $A_0$, which would require a stronger approximation of solutions than the second-moment estimate in Lemma 6.2.","Beyond the paper, because the metric used is the weak metric $d_{\\mathcal{P}(L^2(O))}$ rather than a Wasserstein distance, the result carries no rate in $\\varepsilon$; quantifying a rate would presumably need stronger Lipschitz conditions on $f$ and $\\varrho_k$ and an explicit version of the Gronwall argument.","Beyond the paper, the proof's reliance on the strict gap condition (3.28)-(3.29) suggests a boundary-of-parameters test: at equality in (3.28) the absorbing-set estimate and possibly the attractor convergence should fail, giving a sharp threshold for the phenomenon."],"forward_implications":["For every small $\\varepsilon$, the long-time statistical behavior is concentrated on a compact family of probability laws, so empirical or numerical ensembles of solutions converge uniformly over all translations of the almost-periodic forcing.","The attractor decomposes as a union of kernel sections; consequently each long-time limit can be embedded in a complete bounded trajectory of the law-evolution process.","As $\\varepsilon\\to 0$, no part of the thin-domain attractor escapes to infinity in law space: every weak limit point of $A_\\varepsilon$ lies in the embedded limit attractor $A_0\\circ I^{-1}$.","Uniform tail estimates are strong enough to make the absorbing set and attractor independent of the forcing translation in $H(g_0)$, so the result is genuinely uniform in the non-autonomous symbol."],"supporting_citations":[{"why":"Supplies the existence and uniqueness criterion for uniform measure attractors (Theorem 2.2) that the paper applies to the law-evolution process.","marker":"[28]"},{"why":"Provides the well-posedness result for McKean-Vlasov stochastic reaction-diffusion equations on unbounded domains, giving the unique solutions used to define $U^{g,\\varepsilon}$.","marker":"[9]"},{"why":"Supplies the inequality relating the nonlinearity to $A_\\varepsilon$ that yields uniform $H^1_\\varepsilon$ bounds independent of $\\varepsilon$.","marker":"[27]"},{"why":"Supplies the average operator $M$ and the estimate comparing a thin-domain function with its horizontal average, used in the $\\varepsilon\\to0$ comparison.","marker":"[15]"},{"why":"Provides the abstract semigroup and global-attractor framework on the extended phase space from which the uniform measure attractor theory is drawn.","marker":"[5]"},{"why":"Provides Bochner's criterion and the almost-periodic hull $H(g_0)$, defining the family of external forcings over which all estimates are uniform.","marker":"[26]"}],"fun_headline_variants":["Measure attractors on thin domains: existence and continuity at collapse","Thin-domain stochastic systems: measure attractors converge on collapse","As thin domain collapses, measure attractors exist and converge","Uniform measure attractors survive thin-domain flattening"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The existence proof applies Theorem 2.2, whose hypotheses require joint continuity of the process family on the whole space $\\mathcal{P}_4(L^2(O))\\times H(g_0)$, but the paper proves continuity only on bounded subsets of that space; the unproved full-space continuity is the load-bearing premise.","fun_headline_variants_meta":{"raw":{"variants":["Measure attractors on thin domains: existence and continuity at collapse","Thin-domain stochastic systems: measure attractors converge on collapse","As thin domain collapses, measure attractors exist and converge","Uniform measure attractors survive thin-domain flattening"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001523,"raw_usage":{"total_tokens":6089,"prompt_tokens":921,"completion_tokens":5168,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":537,"completion_tokens_details":{"reasoning_tokens":5101}},"tokens_in":537,"tokens_out":5168,"duration_ms":31706,"temperature":1.0,"reasoning_tokens":5101,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T00:36:32.812301+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete observation that would settle the central claim is a sequence of initial laws $\\mu_n$ in $\\mathcal{P}_4(L^2(O))$ with uniformly bounded fourth moments and forcings $g_n\\to g$ in $H(g_0)$ whose solution laws at times $t_n\\to\\infty$ admit no weakly convergent subsequence; that would falsify uniform asymptotic compactness and with it Theorem 5.1. Alternatively, a sequence $\\varepsilon_k\\to0$ and $\\mu_k\\in A_{\\varepsilon_k}$ whose laws do not converge along any subsequence to an element of $A_0\\circ I^{-1}$ would falsify the upper-semicontinuity claim of Theorem 6.1.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the existence and uniqueness criterion for uniform measure attractors (Theorem 2.2) that the paper applies to the law-evolution process."},{"cited_title":"Chen and B","cited_arxiv_id":null,"evidence_quote":"Provides the well-posedness result for McKean-Vlasov stochastic reaction-diffusion equations on unbounded domains, giving the unique solutions used to define $U^{g,\\varepsilon}$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the inequality relating the nonlinearity to $A_\\varepsilon$ that yields uniform $H^1_\\varepsilon$ bounds independent of $\\varepsilon$."},{"cited_title":"Hale and G","cited_arxiv_id":null,"evidence_quote":"Supplies the average operator $M$ and the estimate comparing a thin-domain function with its horizontal average, used in the $\\varepsilon\\to0$ comparison."},{"cited_title":"Chepyzhov and M","cited_arxiv_id":null,"evidence_quote":"Provides the abstract semigroup and global-attractor framework on the extended phase space from which the uniform measure attractor theory is drawn."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides Bochner's criterion and the almost-periodic hull $H(g_0)$, defining the family of external forcings over which all estimates are uniform."}],"review_version":1}